# Overview Guides and workflows for the Anara research workspace. Anara is an AI research workspace where agents complete research tasks end to end. Give it a goal and Anara searches your library, the web, and academic databases, then synthesizes what it finds into a cited, verifiable answer or document you can edit. The workspace has three parts that work together: a **library** for your papers, documents, and notes; an **agent** that searches, reads, and cites them; and an **editor** for writing with your sources beside you. The guides below teach each part. The workflow guides show how to use them together on research tasks. **Building with the API?** Head to the [API docs](/api) for search, API keys, and integration guides. ## Start here [#start-here] From sign-up to your first cited answer. Uploads, DOIs, web pages, and connectors. File, folder, and general research conversations. Library, web, connectors, and @mentions. A column per question, a row per file. Draft, cite, and export from the note editor. ## Research workflows [#research-workflows] * [Set up a library](/guides/library-setup): collect and check the sources for a project. * [Find literature](/guides/literature-search): search for papers and decide what to import. * [Screen and extract](/guides/systematic-reviews): apply review criteria and extract evidence with Views. * [Transcribe recordings](/guides/transcription): work from audio, video, and YouTube content. * [Check claims](/guides/claim-verification): compare a draft with its supporting sources. * [Create outputs](/guides/agent-outputs): prepare documents, decks, analysis, and charts. ## More resources [#more-resources] * **API:** paper lookup and literature search for your own applications. See the [API tab](/api). * **Account and support:** [account security](/guides/account-security), [plans and billing](/guides/plans-and-billing), [usage and credits](/guides/usage-and-credits), and [troubleshooting](/guides/troubleshooting). * **Open Anara:** start a research task at [anara.com/new](https://anara.com/new). ## For agents [#for-agents] Every page is available as Markdown. Append `.md` to any docs URL, or send `Accept: text/markdown`. [llms.txt](/llms.txt) lists every page and [llms-full.txt](/llms-full.txt) has the full text. An MCP server with `list_docs`, `search_docs`, and `read_doc` tools listens at `https://docs.anara.com/mcp`. See [Use these docs with AI](/guides/docs-for-ai). --- # Quickstart Go from sign-up to your first cited answer in a few minutes. Anara is a research workspace: a library for your sources, an agent that searches and reads them, and an editor for writing. This page gets you from an empty account to a cited answer. ## 1. Create your account [#1-create-your-account] Go to [anara.com](https://anara.com) and sign up with Google, Apple, Microsoft, or your email address. Anara creates a workspace for you automatically. ## 2. Add a document [#2-add-a-document] 1. In the sidebar, click **Create > Upload file** (or **Upload folder**), or drag a file anywhere into the Anara window. 2. Wait a moment while Anara extracts the text and metadata and writes a short summary. Already use Zotero or Mendeley? [Connect it](/guides/connectors) instead of uploading papers one by one. ## 3. Ask a question [#3-ask-a-question] 1. Open the file. 2. Ask a question in the chat on the right and press Enter. 3. Click any numbered citation in the answer to jump to the exact passage it came from. ## 4. Ask across many files [#4-ask-across-many-files] Put related papers in a folder, open the folder, and ask your question there. Ask Anara to search the folder, including subfolders, and cite the sources it uses. For a question that must be answered for each file, use [Views](/guides/views). ## 5. Write a note [#5-write-a-note] Click **Create** in the left sidebar and choose **Note**. Write with the agent beside you, insert citations from your library, and export to PDF, Word, Markdown, HTML, or LaTeX when you are done. ## Next steps [#next-steps] Uploads, DOIs, web pages, and connectors. File, folder, and general research conversations. Library, web, connectors, and @mentions. Build a project from uploads and discovered papers. --- # The interface Find your way around the library, tabs, chat, and file details. Use the left sidebar to navigate, the main area to work with files, and the right sidebar for chat and source details. ## Navigation and Home [#navigation-and-home] The workspace name opens the workspace switcher and settings. **Create** starts new content. **Library** shows your files, **Chats** opens past conversations, and **Search** opens the [command menu](/guides/keyboard-shortcuts). Home includes quick actions and recent files. **Customize home sections** lets you reorder or hide its sections. Recent files may not appear until you have started using the workspace. ## Library [#library] The library table shows files with metadata such as title, authors, date added, full-text status, type, summary, and tags. Use **Filter**, **Sort**, and **Display** to control the table. Select rows for bulk actions. See [organize your library](/guides/organize-library). ## Tabs and split views [#tabs-and-split-views] Files, notes, and chats open in tabs. Drag a tab to a side of the main area to arrange views next to each other. This is useful for reading two papers or keeping a note beside a source. Use the chat-panel toggle when you need more room. ## Right sidebar [#right-sidebar] The available tabs depend on the item you have open. | Tab | Purpose | | ----------- | ------------------------------------------------------------- | | Chat | Ask about the file or folder | | Notes | Work with notes linked to a file | | Details | Inspect metadata, summary, abstract, and citation information | | Annotations | Review PDF highlights and comments | For a note's saved versions, open its menu and choose **Version history**. It is not a separate History tab. Use **Usage** to check AI consumption and **Support** for help. Deleted items are available through **Trash** for the retention period shown there. --- # Set up your library Build a project library from files, connectors, and academic search. Start with a project folder and the sources you already know you need. You can add local files, import from a reference manager, or ask Anara to find more papers. ## Build a project [#build-a-project] 1. Choose **Create > Folder** and give it a recognizable project name. 2. [Upload files](/guides/add-files) or connect [Zotero](/guides/zotero), [Mendeley](/guides/mendeley), or [cloud storage](/guides/connectors). 3. Move the relevant files into your project folder. 4. Check **Full text** and correct important metadata in **Details**. 5. Add tags for review status or methodology. ## Find missing sources [#find-missing-sources] ```text title="Prompt" I'm starting a review of school-based mental health interventions for adolescents. Find relevant empirical studies from the last ten years. Show titles, authors, years, and why each study fits. Let me choose what to import into my project folder. ``` Review the candidates, then ask Anara to import your selection. A source found by search may only provide an abstract. Upload the full PDF through your authorized access when needed. ## Ask your first cross-file question [#ask-your-first-cross-file-question] Open the project folder and use its chat. For example: "Which outcome measures do these studies share? Cite the passages and say when a measure is not reported." Use [folder instructions](/guides/customize#folder-instructions) for project-wide rules and [Views](/guides/views) when you need the same fields from every file. For screening and extraction, follow the [systematic review guide](/guides/systematic-reviews). --- # Keyboard shortcuts and command menu Search, navigate, and edit without leaving the keyboard. Press **Cmd+K** on Mac or **Ctrl+K** on Windows to open Anara's command menu. Search for files, chats, or settings, create a note, upload a file, or paste a URL or paper identifier to import it. Press **Tab** in the menu to ask AI about your input. ## Navigation and chat [#navigation-and-chat] | Action | Mac | Windows | | ------------------- | -------------- | ---------------- | | Open search | Cmd+K | Ctrl+K | | Find in view | Cmd+F | Ctrl+F | | Toggle left sidebar | Cmd+\\ | Ctrl+\\ | | Toggle chat panel | Cmd+/ | Ctrl+/ | | Back / forward | Cmd+\[ / Cmd+] | Ctrl+\[ / Ctrl+] | | Dismiss a menu | Escape | Escape | | Send a chat message | Enter | Enter | | New line in chat | Shift+Enter | Shift+Enter | In the Mac desktop app, **Cmd+,** opens settings, **Cmd+1** through **Cmd+8** select tabs, and **Cmd+9** selects the last tab. ## Writing [#writing] | Action | Mac | Windows | | ------------------------- | ------------------- | -------------------- | | Bold / italic / underline | Cmd+B / I / U | Ctrl+B / I / U | | Inline code | Cmd+E | Ctrl+E | | Strikethrough | Cmd+Shift+X | Ctrl+Shift+X | | Heading 1 through 6 | Cmd+Alt+1 through 6 | Ctrl+Alt+1 through 6 | | Bulleted / numbered list | Cmd+Shift+8 / 7 | Ctrl+Shift+8 / 7 | | To-do list | Cmd+Shift+9 | Ctrl+Shift+9 | | AI assistant | Cmd+J | Ctrl+J | Shortcuts apply to the focused view. When a media player is focused, Space plays or pauses, and the left and right arrows move back or forward five seconds. Browser or operating-system shortcuts can take precedence. --- # Add files Bring files, web pages, papers, and recordings into your library. Add sources to your library so you can organize them, read them, and ask questions with evidence from their content. ## Choose an import method [#choose-an-import-method] | Method | How | | ------------------------ | -------------------------------------------------------------------------------------------------------- | | Local files | Drag files into Anara or use **Create > Upload file** | | Local folder | Use **Create > Upload folder** | | Link or paper identifier | Paste it into [Search](/guides/import-by-url) | | YouTube | Import the video URL | | Web page | Import the URL or use the [Chrome extension](/guides/chrome-extension) | | Connected service | Use a [connector](/guides/connectors) and select what to import | | Academic search | Ask the agent to [find papers](/guides/literature-search), review the results, and import your selection | See [supported file types](/guides/file-types) for the full format list and file limits. Importing uses [AI usage](/guides/usage-and-credits). ## Check processing [#check-processing] Anara extracts available text and metadata, then indexes readable content for search. Check the **Full text** column after processing. | Status | Meaning | Next step | | ------------------- | ------------------------------------------------- | ----------------------------------------------------------- | | Available | Readable content was extracted | Open the file and check what you need | | Abstract only | The abstract is available, but not the full paper | Upload the full PDF through your authorized access | | Content unavailable | Readable content could not be obtained | Check the failure and try the relevant troubleshooting step | Metadata can be incomplete or wrong. Correct important fields in **Details** before relying on citations. ## Scans, images, and figures [#scans-images-and-figures] Scanned files can need OCR, which turns page images into searchable text. Anara's OCR fallback is available on Pro and above. You can also upload a searchable PDF with an existing text layer. The agent can inspect available figures as well as read extracted passages. Ask it to examine the particular figure or page, then check labels and numerical values against the original. Extracted text alone can miss visual information. For failed, locked, or incomplete files, see [troubleshooting](/guides/troubleshooting). --- # Supported file types Formats you can upload and the limits that apply to them. Upload files from your computer or drag them into Anara. These are the supported upload extensions. | Content | Extensions | | ------------------ | ------------------------------------------------------------------ | | PDF | `.pdf` | | Word and rich text | `.doc`, `.docx`, `.rtf` | | Presentations | `.ppt`, `.pptx` | | Spreadsheets | `.xls`, `.xlsx`, `.xlsm`, `.ods`, `.csv` | | Text and ebooks | `.md`, `.txt`, `.epub` | | Images | `.png`, `.jpg`, `.jpeg`, `.tiff`, `.bmp`, `.heic`, `.gif`, `.webp` | | Audio | `.mp3`, `.m4a`, `.wav` | | Video | `.mp4` | [Web pages and paper identifiers](/guides/import-by-url) use link import. [Audio and video](/guides/transcription) are transcribed so you can ask about what was said. ## File limits [#file-limits] | Plan | File size limit | Page limit | | ------------------ | --------------- | ---------- | | Free | 20 MiB | 120 | | Plus and Education | 100 MiB | 600 | | Pro | 300 MiB | 10,000 | MiB means 1,048,576 bytes. These are the file and page limits, not a daily upload allowance. Importing also uses [AI usage](/guides/usage-and-credits). Markdown and plain-text imports have a separate 5 MiB limit. Max and Enterprise have different entitlements. A plan without a fixed file-size ceiling still depends on successful upload and processing; do not treat it as a guarantee that any file can be processed. ## After upload [#after-upload] The library distinguishes files such as notes, spreadsheets, images, audio, video, and websites. **Full text** indicates whether readable content was extracted. An accepted extension does not guarantee that an encrypted, damaged, or incomplete file can be read. See [troubleshooting](/guides/troubleshooting). --- # Import by link or identifier Save a web page, video, or paper without uploading a local file. Import a link or paper identifier to keep the source in your library. Anara resolves available metadata and retrieves readable content when it can. ## Import a source [#import-a-source] 1. Open **Search**, or press Cmd+K on Mac or Ctrl+K on Windows. 2. Paste a web URL, YouTube URL, DOI, or PMID. 3. Choose the import action offered for that input. 4. Wait for processing, then check the new item in your library. For a DOI, paste the identifier or its `https://doi.org/` link. For PubMed, a PubMed URL identifies the paper without ambiguity. You can also ask chat to find a paper and import the result after you review it. ## Full text and paywalls [#full-text-and-paywalls] **Available** means readable content was extracted. **Abstract only** means Anara has the abstract but not the paper's body. **Content unavailable** means it could not obtain readable content. Importing a DOI does not bypass a publisher's paywall. If you have institutional access, download the PDF yourself and [upload it](/guides/add-files). Confirm that the imported record matches the title, authors, and year you intended. ## Usage and failures [#usage-and-failures] Imports draw on your [usage allowance or credits](/guides/usage-and-credits). A URL that requires a login, has expired, or points to a navigation page may fail to provide the source. Try a direct paper or PDF link, or upload a local copy. For spoken content, see [transcription](/guides/transcription). --- # Organize your library Use folders, tags, metadata, and filters to keep a growing library findable. Folders give your library structure, tags add labels that cut across folders, and metadata powers search, filters, and citations. ## Folders [#folders] 1. Click **Create** in the left sidebar and choose **Folder**. 2. Right-click a folder and choose **New folder** to nest one inside it. 3. Drag files onto a folder in the sidebar, or select several with checkboxes and move them together. Folders matter for chat as well as tidiness. Agent chat treats an open folder as its current context, so answers draw on that project first instead of your whole library. Each folder is Private, Team, or Shared; see [Share and collaborate](/guides/collaborate). ## Tags [#tags] A file lives in one folder but can carry many tags, such as `RCT`, `to-cite`, or `key-paper`. * Click the **Tags** cell in the library table, type a name, and press Enter. New tags are created as you type them. * Or open a file and add tags in the **Details** tab. Use folders for projects and topics, and tags for status and methodology. ## Metadata [#metadata] Anara fills in title, authors, publication, volume, issue, pages, DOI, publication date, citation count, abstract, and an AI summary where it can. Importing by DOI or PMID can retrieve a paper’s bibliographic record. To fix a field, open the file, go to **Details**, and click the field. Check metadata before using it in a formatted citation. ## Find things [#find-things] * **Search** in the left sidebar finds files by title, content, authors, and other metadata. * **Filter** narrows the table by fields such as folder, tags, date added, file type, or full text status. Filters combine. * **Sort** orders by title, date, publication date, citation count, and more. * **Display** shows or hides columns. Filters and sorting apply to whatever is selected in the sidebar, so click **Library** first to work across everything. ## Trash [#trash] Deleted files, chats, and folders go to **Trash** in the left sidebar and can be restored during the retention period shown in Trash. ## Library setup and file details [#library-setup-and-file-details] For a new project, follow [library setup](/guides/library-setup). Use **Details** to inspect the file metadata, abstract, and summary, then correct the editable fields. Filters include bibliographic fields such as title, authors, publication, DOI, ISSN, ISBN, URL, abstract, volume, issue, pages, and citation count; organizational fields such as folder, visibility, and tags; and added date, publication date, file type, full text, and summary. Available columns and filters depend on the view. --- # Transcribe audio and video Import recordings and ask questions grounded in their transcripts. Upload audio or video to make spoken content available for research. You can also import a YouTube URL. ## Add a recording [#add-a-recording] 1. Upload an MP3, M4A, WAV, or MP4 file, or [import a YouTube link](/guides/import-by-url). 2. Wait for processing and transcription. 3. Open the audio or video item. 4. Ask about the recording and follow its citations back to the relevant content. ```text title="Prompt" Summarize this lecture's main arguments. For each one, give the supporting transcript passage and timestamp where available. Separate the speaker's claims from your interpretation. ``` ## Review a transcript [#review-a-transcript] Transcription can mishear names, numbers, abbreviations, and overlapping speech. Listen to the original segment before using an exact quotation or consequential detail. Tell the agent which term was misheard when asking a follow-up. You can organize recordings into folders alongside papers, ask comparative questions, or use [Views](/guides/views) for consistent extraction across supported recordings. ## If processing fails [#if-processing-fails] Check the [format and file limits](/guides/file-types), your connection, and available [usage](/guides/usage-and-credits). A private, unavailable, or restricted video URL may not be retrievable. Upload an authorized local recording when link import cannot access it. --- # Extract with views Turn a folder into a table where every file answers the same questions, with citations. A view is a table built from a folder. Each row is a file, each column is a question, and Anara extracts answers from each file. Cells with source evidence let you open the supporting passages and check the result. Use a view when you want the same answer from every file: sample sizes, study designs, outcomes, limitations, or the fields you need to screen papers. For one synthesized answer across a folder, use [chat](/guides/chat) instead. ## Create a view [#create-a-view] 1. Open a folder. 2. Click **View** in the toolbar, then **Create view**. 3. Describe what to extract, or click a suggested prompt and edit it. 4. Click the arrow. Anara writes the columns, names the view, and fills cells column by column. To use only some files, select them with checkboxes and click **Extract to table**. You can also ask in a folder chat, for example: "create a table extracting the country of the first author and the funding source from this folder." Describe the fields, not the task. "Extract sample size, study design, primary outcome, and limitations" gives cleaner columns than "review these papers". ## Edit columns [#edit-columns] * Click **Add extraction column**, write a question such as "What was the sample size?", and pick a type: Text, Number, Date, Yes / no, or Single select. * Open a column header to edit, retry, or delete it. Changes to extraction instructions can make existing values stale. Rerun the column to update them. * Ask the chat to add, change, rerun, rename, or delete columns in the open view. ## Check the answers [#check-the-answers] Hover a cell and click the passage count to open **Source evidence**: the file, page, and exact text used. Click a row to open the file at that passage. | Cell shows | Meaning | | ------------ | ------------------------------------------------------------------------------ | | Not reported | Anara did not find an answer. Check the source and reword the column if needed | | Failed | The extraction failed. Check the file and error, then retry | | Waiting | The cell is queued for extraction | ## Keep it current [#keep-it-current] A view built from a whole folder follows it: new files add rows, removed files drop them. Click **Rerun** to refresh every cell. Views read documents, web pages, spreadsheets, notes, video, and audio; images are skipped. One run fills up to 2,000 cells (files multiplied by columns). --- # Chat with Anara Ask about a file, a folder, or a research question and check the sources. Anara can search sources, read relevant content, and use tools to complete a research task. Tell it what you need and check the evidence behind its answer. ## Start a conversation [#start-a-conversation] Use **Agent** or **Create > Agent** for a general conversation. Open a file or folder and use **Chat** in the right sidebar for a conversation associated with that item. The sidebar label is Chat. Ask a focused first question, such as "What are the methods and limitations in this paper?" You can use a suggested starter question if one fits. For a folder, ask a question across its files and specify whether external sources are allowed. ## What Anara can do [#what-anara-can-do] * Search your library, the web, academic literature, and connected services. * Read source passages and available full pages, URLs, and media. * Create or edit notes, spreadsheets, files, charts, and images. * Import sources and organize workspace content. * Plan a multi-step task and ask for clarification when needed. Review any proposed import, edit, or external action before approving it. For a larger task, ask for a plan first and state the output you want. See [workspace actions and outputs](/guides/agent-outputs). ## Ask about a passage [#ask-about-a-passage] Select text in a PDF and choose **Chat** to include the passage in your question. Use **Explain** for a focused explanation of selected text. You can also mention a file, folder, or source with `@`. ## Continue and check [#continue-and-check] Chats save so you can return to them. Use **Chats** or [Cmd+K search](/guides/keyboard-shortcuts) to find a conversation. Use **Stop** to interrupt a response and the available retry or regenerate control to try again. Open citations and read their surrounding context. If an answer is wrong, identify the unsupported claim and ask for a correction from the source. For recurring instructions, use [Skills and folder instructions](/guides/customize). --- # Sources and @mentions Choose the library, web, files, folders, and connected apps for a question. Source controls tell Anara where it may look. Mentions identify sources or actions for a particular request. ## Source controls [#source-controls] Open the source controls in the chat composer. Depending on your interface, these are in the **+** menu or beside the input. * Disable **Web** when the answer should come from your own material. * Disable **Library** when you do not want a library-wide search. * Open or mention the specific file or folder when the task has a defined evidence set. State the boundary in the prompt as well: "Use only the attached file and say when it does not answer the question." A source setting does not guarantee that an answer is complete or correct. Check its citations. ## Mention a source or tool [#mention-a-source-or-tool] Type `@` and search the menu. It includes library and web sources, connected apps, files and folders, and actions such as **Create note**, **Edit note**, **Create citation**, **Create image**, and **Plan**. Availability depends on your connections and plan. ```text title="Prompt" @Library Compare the study designs in my project folder. Cite the relevant methods passages and flag files that only have abstracts. ``` Connect an app before using it as a source. See [connectors](/guides/connectors). For academic research outside your library, see [literature search](/guides/literature-search). ## Notes and project rules [#notes-and-project-rules] **Settings > Preferences > Include notes in AI responses** controls whether your notes can be read and cited. Use [folder instructions](/guides/customize#folder-instructions) to keep standing project rules with a folder. --- # Find academic literature Discover papers, check matches, and build a source library. Ask Anara to find papers by research question, title, author, or identifier. State the field, dates, and study types you need so it can choose useful searches. ## Search by topic [#search-by-topic] ```text title="Prompt" Find empirical studies on this research question published since 2020. Give me the title, authors, year, study design, and why each paper is relevant. Separate reviews and preprints from primary studies. ``` Use **Web** as a source for external research. Connect a [science database](/guides/science-databases) when you need to name a particular database. You can combine external search with your own [library sources](/guides/sources). ## Find a known paper [#find-a-known-paper] Provide its DOI or a complete title with author and year. Check the match before importing, especially for common titles or different editions. Use [import by link or identifier](/guides/import-by-url) to save a known source directly. ## Read before drawing conclusions [#read-before-drawing-conclusions] Search results can contain bibliographic metadata and abstracts without the full paper. Ask Anara which sources it read in full. Open the cited papers and check whether the evidence supports the answer. For connected work around a seed paper, use the [citation graph](/guides/citation-graph). For a structured review, record your search question, databases, dates, and criteria, then use [systematic reviews](/guides/systematic-reviews). --- # Models and Deep Search Choose a model and use deeper research modes when the task needs them. Choose a model from the chat composer or set your default under **Settings > Preferences > AI model**. There is one AI-model preference, not separate Chat and Editor defaults. ## Model access [#model-access] Free and Plus use **Auto**. Pro and higher plans with full-model access can choose from the available picker. Switching models applies to the next response; it does not rewrite earlier messages. The picker shows the current model names and relative cost. Model availability can differ for HIPAA-enabled workspaces. Use Auto for routine work and choose another available model when the task calls for it. ## Where models are hosted [#where-models-are-hosted] A model's developer and its hosting provider are different things. The picker marks applicable models with **Available on US-hosted servers**. This includes listed models developed by Zhipu, Moonshot AI, DeepSeek, MiniMax, Tencent, and Qwen. Anara routes these through hosting providers including Baseten, Novita, and Modal. The picker notice describes hosting; it does not mean the model was developed in the United States. Check [Anara security](https://anara.com/security) for service-level data handling and ask support about requirements specific to your workspace. ## Model Council [#model-council] On Max, Model Council asks multiple models to work on a question and synthesizes their answers. Use it when comparing approaches is useful. Review the underlying evidence even when the models agree. It uses more AI work than a single-model answer. ## Deep Search [#deep-search] Deep Search is available on Max plans. Open the chat composer's **+** menu and enable **Deep search**. It applies to subsequent requests until you turn it off. Use it for multi-step research, comparisons across sources, and questions that need a more extensive investigation. Name the project folder and source boundaries first. For a short factual question, ordinary chat may be sufficient. Deep Search and Model Council draw on your [usage and credits](/guides/usage-and-credits). They do not make an incomplete source set exhaustive. For structured study screening, use [Views](/guides/systematic-reviews). --- # How citations work Follow evidence back to its source and choose a consistent format. Citations connect an answer to its sources. Open them to check whether the evidence supports the statement, including any qualifications around the cited passage. ## Check a citation [#check-a-citation] A library citation can open the file at the relevant passage. A web or database citation links to the external source. Source availability varies: a paper with only an abstract cannot support claims about methods or results that are absent from that abstract. Ask Anara to distinguish the source's statements from its own inference. A citation beside a sentence does not by itself establish that the sentence is correct. For a draft-level check, use [claim verification](/guides/claim-verification). ## Choose a format [#choose-a-format] Open **Settings > Preferences > Citation format**. Search the catalog by style or journal name, or [upload or generate a custom style](/guides/citation-styles). This is one global setting across chats, notes, and exports. The citation format in the note menu changes the same preference. Correct source metadata in **Details** when authors, dates, or identifiers are missing or wrong. ## Cite your writing [#cite-your-writing] In a note, select a claim and choose **Search citation**. Preview candidates and insert the source that actually supports the text. See [write notes](/guides/notes). ## Keep the references [#keep-the-references] Open an answer's source list to inspect its metadata or formatted references. Copy or download it as TXT, BibTeX, RIS, or CSV. See [reference lists](/guides/reference-lists) for the full workflow. --- # Trace a paper’s citations Find references, citing papers, and related work around a seed paper. Use a paper as the starting point for finding connected research. Ask for papers it cites, papers that cite it, or related work. ## Start with a paper [#start-with-a-paper] Provide the DOI, or the title with author and year. Ask Anara to confirm the paper before tracing its connections. ```text title="Prompt" Find this paper by DOI: [paste DOI]. Show the papers that cite it and explain which ones test, extend, or challenge its findings. Give me source links and distinguish what you read from what you inferred from an abstract. ``` To search backward, ask for the references in the seed paper's bibliography. To expand by topic, ask for related papers and explain what makes a result relevant. ## Check coverage [#check-coverage] Citation data depends on the available index. A missing connection is not evidence that a paper has never been cited. A highly cited paper is not necessarily the best source for your question, and a citation does not imply agreement. Open relevant results, read the evidence, and [import your selection](/guides/import-by-url) into the project. Use [claim verification](/guides/claim-verification) to check what a source actually supports. --- # Verify claims against sources Check your draft one claim at a time and inspect the supporting passages. Ask Anara to compare a claim with the evidence in your files. Treat the output as a review you can inspect, not a guarantee that the claim is correct. ## Check a draft [#check-a-draft] 1. Upload the sources or open the project folder containing them. 2. Paste the exact sentence or paragraph to check. 3. State whether external sources are allowed. 4. Ask for a verdict and evidence for each claim. ```text title="Prompt" Check the paragraph below against the files in this folder. Break it into claims. Mark each as supported, partially supported, contradicted, or not found. Quote the relevant passages with citations and identify missing full text. Do not fill gaps from general knowledge. [paste paragraph] ``` ## Review the evidence [#review-the-evidence] Open each citation and read the surrounding passage. Check population, methods, dates, and qualifications. A passage on the same subject may not support the sentence as written. "Not found" means the available search did not establish support; it does not prove the statement false. Ask for a revision that stays within the evidence, then check it again. In a note, [Search citation](/guides/notes#add-citations) helps you find and insert a source for selected text. --- # Screen and extract for a systematic review Use a project library and Views to apply consistent questions to each study. Use Views to ask the same screening or extraction questions of every file in a project. Keep your protocol and human decisions explicit so you can review how the evidence was selected. ## Prepare the library [#prepare-the-library] Write inclusion and exclusion criteria before screening. [Find and import the studies](/guides/literature-search), organize them into a folder, and check full-text availability. Use tags to track stages such as awaiting full text, included, or excluded. ## Screen with a view [#screen-with-a-view] Open the folder and [create a view](/guides/views). Ask for separate columns for the criteria, a proposed decision, and the reason. For example: ```text title="Extraction request" For each study, extract population, study design, intervention, comparator, and publication year. Propose Include, Exclude, or Unclear against the criteria below. Give the reason and supporting passages. Use Unclear when the available text is insufficient. ``` Review the supporting evidence in each row and record the final decision. Do not treat an AI screening result as an independently completed human review. ## Extract from included studies [#extract-from-included-studies] Create columns for the fields your protocol needs, such as sample size, outcomes, time points, effect estimates, uncertainty, and limitations. For PICO, keep population, intervention, comparator, and outcome in separate columns. Ask for "not reported" rather than an inferred value. Inspect **Source evidence** for important cells. Keep units and follow-up periods explicit before comparing studies. Use a [spreadsheet](/guides/spreadsheets) for calculations and analysis once the extracted values have been checked. Views support this workflow but do not by themselves establish that a search is exhaustive or a review meets a reporting standard. --- # Skills and folder instructions Reuse methods and keep standing rules with a project. Use Skills for reusable methods and folder instructions for rules that belong to a project. ## Skills [#skills] Type `/` in chat to find an available Skill. A Skill gives the agent instructions for a repeatable task. Tell it which files to use and what result you want. Open **Settings > Skills** to manage the Skills available to you. Start with an existing method where it fits and adapt your instructions for the work at hand. Avoid relying on an instruction to compensate for missing source material. ```text title="Reusable task" For each study, identify its research question, population, methods, results, and limitations. Cite the passages and mark anything not reported. ``` ## Folder instructions [#folder-instructions] Open the folder's settings and choose **Instructions**. Enter the standing rules you want Anara to follow in chats in that folder. People with permission to modify the folder can change these instructions. Examples: * Answer from this project's files and state when evidence is missing. * Report units and uncertainty with quantitative results. * Use the terminology in our project glossary. Instructions are saved with the folder. For an isolated request, put the rule directly in the message. For language, model, or citation defaults, use [Preferences](/guides/settings). --- # Write better prompts Ask specific questions, scope them, and plan big tasks. Describe the result you need and the sources you want Anara to use. Clear instructions help it search for relevant evidence and shape the answer. ## Be specific [#be-specific] | Instead of | Ask | | ------------------------ | -------------------------------------------------------------------------------- | | Tell me about this paper | What statistical methods did the authors use, and what were their limitations? | | Summarize my research | What are the three most common themes across the papers in my climate folder? | | Papers on cancer | Papers on PD-L1 inhibitors in triple-negative breast cancer published after 2022 | Say what you will do with the answer, such as a table for a systematic review or a paragraph for a methods section, and the format you want back. ## Scope the question [#scope-the-question] * Type `@` and pick a file or folder to point the agent at the sources you care about. See [Sources and @mentions](/guides/sources). * Turn off web search when the answer must come from your documents. * Select a passage in a PDF and send it to chat, rather than describing where it is. ## Plan big tasks [#plan-big-tasks] For multi-step work such as a systematic review, type `@Plan` and ask the agent to outline its approach first. You can correct the plan before it starts, which catches misunderstandings early. ## Reuse what works [#reuse-what-works] To reuse a Skill or set rules for a folder, see [Skills and instructions](/guides/customize). ## Control length [#control-length] Ask for the length you want in the message itself, such as "a brief summary" or "a detailed analysis". ## Learn from examples [#learn-from-examples] Start with [library setup](/guides/library-setup), [systematic reviews](/guides/systematic-reviews), or [claim verification](/guides/claim-verification) for prompts you can adapt. --- # Workspace actions and outputs Create files, analyze data, and organize research from chat. Tell Anara the result you need: a note, a comparison table, a chart, a document, or a set of organized sources. Include the input files and the format you want. ## Documents and decks [#documents-and-decks] Ask for a Word document or PowerPoint presentation from your checked sources. Specify the audience, length, structure, and citation requirements. Download the attached result and inspect its layout before sharing it. ```text title="Prompt" Create a short presentation comparing the methods in these papers. Include a source for each factual slide and a final slide listing unresolved questions. ``` ## Analysis, charts, and spreadsheets [#analysis-charts-and-spreadsheets] The agent can run code for calculations and file creation. Ask it to explain its assumptions, check missing values, and show the source data used in a chart. It can create and edit [spreadsheets](/guides/spreadsheets) as well as generate downloadable files. For an animation or interactive output, state the behavior and desired file format. Inspect the output in the tool where you plan to use it; a generated file still needs validation. ## Notes and library actions [#notes-and-library-actions] Ask Anara to create or edit a note, import a paper, create a folder, move selected files, or update metadata. Name the exact items and destination. Review approval requests for the proposed action before continuing. Use `@Create note` or `@Edit note` to make the intended action explicit. For repeatable methods, use [Skills](/guides/customize). ## Images [#images] Ask for an image or attach an image and describe the edit. `@Create image` selects the image action. Specify labels, layout, and what must remain accurate. Check generated scientific illustrations against your sources. ## Research execution [#research-execution] Anara can read accessible URLs and split larger work into smaller research tasks. Ask for a plan and a source-backed result rather than depending on a particular internal execution sequence. ## Export [#export] Use an item's [Export menu](/guides/export) for supported note and chat formats. Use the source list for [reference exports](/guides/reference-lists). Downloadable agent-created files are separate from these menu exports. --- # Read and annotate Read PDFs, inspect figures, and keep highlights and comments. Open a PDF from your library to read it alongside chat. Use the viewer's page and zoom controls and **Find** to locate text. ## Select a passage [#select-a-passage] Select text to open the floating toolbar. Choose **Chat** to include the passage in a question, **Highlight** to save it, or **Comment** to attach a note to the highlight. Use **Explain** for a focused explanation of selected text in a PDF or chat answer. Click a highlight to review or change it. The **Annotations** tab lists highlights and comments; choose one to return to its passage. ## Read figures and scanned pages [#read-figures-and-scanned-pages] Ask Anara about a specific figure or page when the information is visual. Check axes, units, legends, and numerical values against the original. A text search alone does not establish what a figure shows. See [OCR and processing](/guides/add-files#scans-images-and-figures). ## File details [#file-details] Open **Details** for metadata, the abstract, an AI summary, and citation information. Edit incorrect fields where the interface allows it. A summary is a starting point for reading, not a substitute for the source. ## Compare sources [#compare-sources] Open files in separate tabs and [arrange them side by side](/guides/interface#tabs-and-split-views). Ask chat to compare their methods or conclusions, with citations to both. ## Download annotations [#download-annotations] Open the PDF menu and choose **Download with annotations** to keep a copy containing your highlights and comments. The option appears when annotations exist. See [export](/guides/export). If Anara cannot find content you can see in the file, check **Full text** and the [processing guidance](/guides/troubleshooting). --- # Write notes Draft, revise, cite, and export with your sources beside you. Notes are files you write in Anara. Keep them in project folders beside the sources that support your work. ## Create and format [#create-and-format] Choose **Create > Note**, create one from the command menu, or ask the agent to create a note. Use the editor toolbar to format headings, lists, to-do lists, tables, quotes, code blocks, and links. Select text to reveal contextual editing actions. The editor supports LaTeX math. Use math blocks or the supported LaTeX syntax for equations and check the rendered result. See [keyboard shortcuts](/guides/keyboard-shortcuts) for common formatting commands. ## Write with the agent [#write-with-the-agent] Open **Chat** beside the note. Ask for a draft or revision with a clear source boundary and purpose. For example: "Draft a limitations section from the papers in this folder. Mark gaps rather than inventing evidence." For **Quick Edit**, select the passage and describe a change such as "make this more concise." Review the proposed result and its citations before keeping it. ## Add citations [#add-citations] 1. Select the claim that needs a source. 2. Choose **Search citation**. 3. Preview the candidate sources. 4. Insert the source that supports the claim. The format follows the global **Settings > Preferences > Citation format** preference. See [citation styles](/guides/citation-styles). ## Collaborate and restore [#collaborate-and-restore] Notes in shared content support collaborative editing when people have write access. To inspect an earlier saved version, open the note menu and choose **Version history**, preview a version, then **Restore** if needed. See [version history](/guides/version-history). ## Export [#export] Open the note menu and choose **Export**. Formats include PDF, Word, Markdown, HTML, LaTeX PDF, and LaTeX source. Check the exported file before sharing. See [export](/guides/export), or use the [Word add-in](/guides/word-add-in) to work inside Word. --- # Work with spreadsheets Read, edit, analyze, and export workbook data. Open a spreadsheet in Anara to inspect its sheets and ask questions about its data. You can also ask the agent to create or edit a workbook for a research task. ## Import and inspect [#import-and-inspect] [Upload](/guides/add-files) Excel, ODS, or CSV files. Open the result and select the sheet you want to inspect. Check headers, units, formulas, and imported values before analyzing them. Use the spreadsheet controls to find values, sort or filter rows, and adjust the view. You can edit cells and manage sheets when you have write access. Wait for changes to save before exporting or closing a task. ## Ask for analysis [#ask-for-analysis] ```text title="Prompt" Compare outcomes by study group in this workbook. Identify missing values and units first. Show the calculations and create a chart from the checked values. Do not replace missing values with zero. ``` Anara can query workbook data, edit cells, and produce analysis outputs. Review calculations and source ranges, especially when a workbook contains formulas or incomplete imported sheets. ## Export [#export] Choose **Export > Download as CSV** for the active sheet or **Download as Excel (.xlsx)** for the workbook. See [export](/guides/export) for other output types. Use [Views](/guides/views) when each row should extract an answer from a separate source file. Use a spreadsheet when you want to work directly with tabular data or calculate across the values. --- # Choose a citation style Use the style catalog or create a custom style from a file or example. One citation-format setting applies across chats, notes, and exports. Choose the style your journal or institution requires before preparing your final output. ## Choose from the catalog [#choose-from-the-catalog] 1. Open **Settings > Preferences**. 2. Under **AI and agents**, open **Citation format**. 3. Search by style or journal name and select the result. You can also open **Citation format** from a note's menu. It changes the same global preference. Correct missing authors, dates, or identifiers in the source's **Details** before judging its formatted reference. ## Upload a custom style [#upload-a-custom-style] 1. Open the custom citation-style option in Preferences. 2. Choose **Upload .csl** and select your Citation Style Language file. The uploader also accepts XML. 3. Check the rendered preview. 4. Enter a style name and click **Save style**. A CSL file describes how citations and bibliographies are formatted. Use a valid style file rather than renaming a Word or PDF file to `.csl`. The uploader accepts files up to 512 KB. ## Generate from an example [#generate-from-an-example] Paste one reference formatted exactly as you want it, then choose **Generate from example**. Inspect the preview. Use the refinement field and **Revise** to request changes such as initials instead of full names or omission of the DOI. Name and save the style when the preview is right. A generated style needs checking against your actual publication requirements. Test several source types, especially if the example only covers journal articles. See [reference lists](/guides/reference-lists) and [export](/guides/export). --- # Create and export reference lists Keep the sources behind an answer and export them for your writing tools. Use the sources attached to an answer to build a reference list you can check and reuse. A reference list describes the sources; inline citations show which passage supports a particular statement. ## Review an answer's sources [#review-an-answers-sources] Open the sources for an answer. Switch between **Metadata** and **Reference** to inspect the source details or formatted references. Filter to **All**, **Library**, or **Web** when you only need one group. Check titles, author names, dates, and identifiers before exporting. For library files, correct missing or wrong fields in **Details**. Choose your [citation style](/guides/citation-styles) before preparing a formatted list. ## Export [#export] The source list offers: * **Copy to clipboard** for pasting into another tool. * **Download as TXT** for a text reference list. * **Download as BibTeX** for LaTeX and reference managers. * **Download as RIS** for reference managers. * **Download as CSV** for a table of source metadata. The export follows the sources currently shown, including the Library or Web filter. ## Ask for a bibliography [#ask-for-a-bibliography] ```text title="Prompt" Create a reference list for the papers used in this answer. Use my selected citation style and flag missing authors, dates, or DOIs rather than inventing them. ``` For a whole folder, name or mention it and ask Anara to work from those files. Review the resulting list against the folder; a bibliography from one answer contains that answer's sources, not necessarily every file in the project. --- # Restore an earlier note version Preview saved versions and restore the one you need. Version history lets you inspect earlier saved versions of a note before restoring one. 1. Open the note. 2. Open its menu and choose **Version history**. 3. Select a saved version to preview it. The list shows when versions were saved and their authors where available. 4. Choose **Restore this version**, then confirm with **Restore**. Previewing a version does not by itself replace the current note. Exit the preview to return to your current work. If you want a separate copy before restoring, duplicate or [export the note](/guides/export). If the list says no versions have been saved yet, there is no earlier version to restore. If a version fails to load, retry before changing anything. Version history is part of the note menu, not a History tab beside the file. --- # Export your work Download notes, chats, annotated PDFs, spreadsheets, and references. Export the item you want to keep from its menu. The available formats depend on the kind of item. | Item | Where | Formats | | ---------------- | ------------------------- | ---------------------------------------------------------- | | Note | Note menu > Export | PDF, Word, Markdown, HTML, LaTeX PDF, LaTeX source (.zip) | | Chat | Chat menu > Export | PDF, Word, Markdown, HTML, LaTeX PDF, LaTeX source (.zip) | | PDF | File menu | Download; Download with annotations when annotations exist | | Spreadsheet | Export in the spreadsheet | CSV for the active sheet; Excel (.xlsx) for the workbook | | Answer's sources | Source-list menu | TXT, BibTeX, RIS, CSV, or copy to clipboard | ## Notes and chats [#notes-and-chats] Choose **Word** for an editable document, **PDF** for a fixed layout, or **Markdown** for plain-text editing. **LaTeX PDF** compiles the output; **LaTeX source (.zip)** gives you source files to continue editing elsewhere. Citation formatting uses your [global citation style](/guides/citation-styles). Open the exported file and check its layout, equations, images, and references before sharing it. ## PDFs with annotations [#pdfs-with-annotations] Open the PDF's menu and choose **Download with annotations** to include your highlights and comments. That option appears when the file has annotations. Use **Download** for the original file. ## Generated files [#generated-files] When the agent creates a document, deck, image, or chart in chat, download the attached result. Ask for the exact format you need. See [what the agent can create](/guides/agent-outputs). ## If export fails [#if-export-fails] Wait until the item has loaded and changes have saved, then retry. If one format fails, another can help preserve your work while you [contact support](/guides/troubleshooting#contact-support). Include the item name and requested format. --- # Where to use Anara Choose the web app, Mac app, iPhone and iPad app, Word add-in, or Chrome extension. Your Anara account connects your research across devices. Sign in to the same account and select the same workspace to find your library. | App | Use it for | | -------------------------------------------- | ---------------------------------------------------- | | [Web app](https://anara.com) | Your research workspace in a browser | | [Mac desktop](/guides/desktop) | The workspace in a dedicated Mac window | | [iPhone and iPad](/guides/ios) | Reading and asking questions on the move | | [Microsoft Word](/guides/word-add-in) | Research, editing, and citations inside Word | | [Chrome extension](/guides/chrome-extension) | Saving web pages while browsing | | [Anara over MCP](/guides/mcp) | Working with your Anara library from another AI tool | For public academic search and paper lookup in your own application, see the [API reference](/api). ## Start on the web [#start-on-the-web] Open [anara.com](https://anara.com), sign in, and follow the [quickstart](/guides/quickstart). The guides describe the web and Mac workspace unless a page names a different app. Menus and available controls can differ on mobile and in Word. --- # Use Anara on Mac Install the Mac app and open your existing research workspace. The Mac app gives you a dedicated window for your Anara workspace. Use the same account and workspace as the web app to continue your research. ## Install [#install] 1. Open [Anara downloads](https://anara.com/download). 2. Choose the Mac download offered for your computer. 3. Open the downloaded installer and install Anara. 4. Launch Anara and complete sign-in. If sign-in opens your browser, return to Anara when it finishes. ## Work in the app [#work-in-the-app] Upload files, organize your library, read and annotate, chat, and write notes using the same workflows described in these guides. You need a connection for AI requests and synced workspace content. The desktop app also has [keyboard shortcuts](/guides/keyboard-shortcuts) for settings and switching tabs. Use Cmd+, for settings and Cmd+1 through Cmd+8 to select a tab. Cmd+9 selects the last tab. ## If something is missing [#if-something-is-missing] Check the account and workspace first. Signing into a different workspace can show a different library. For sign-in or loading problems, see [troubleshooting](/guides/troubleshooting). --- # Use Anara on iPhone and iPad Install the iOS app and take your research library with you. Anara has an app for iPhone and iPad. Use it to read your research and ask questions without opening a desktop browser. ## Install and sign in [#install-and-sign-in] 1. Open [Anara on the App Store](https://apps.apple.com/us/app/anara/id6800030182). 2. Install the app. 3. Sign in with your Anara account and select the workspace you want to use. Your account can belong to several workspaces. If expected files are missing, check the selected workspace before uploading another copy. ## Continue your research [#continue-your-research] Open a file from your library and use chat to ask about it. Follow citations back to their sources to check the answer. Use the web or Mac app for workflows described here that need controls unavailable in the mobile interface. ## Subscriptions bought through Apple [#subscriptions-bought-through-apple] If you subscribed in the iOS app, Apple manages that subscription. Open **Settings > Billing** on the web and choose **Manage in the App Store**, or manage it through your Apple account. Deleting Anara or deleting your Anara account does not cancel an App Store subscription. See [plans and billing](/guides/plans-and-billing) for cancellation and moving to web billing. --- # Use Anara in Microsoft Word Install the Word add-in and research, edit, and cite from your document. Anara for Word opens a research pane beside your document. Use **Agent** to research and edit, and **Cite** to find and insert references. A free Anara account is enough to start; your plan controls usage and model access. ## Install inside Word [#install-inside-word] 1. Open a document in Word on Windows, Mac, or the web. 2. Open **Home > Add-ins > More Add-ins**. Some Word versions put the store under **Insert > Get Add-ins** or **File > Get Add-ins**. 3. Search the store for **Anara**. 4. Select **Anara for Word** and click **Add**. 5. Open Anara from the **Home** tab and sign in to your Anara account. ## Install from Microsoft AppSource [#install-from-microsoft-appsource] Find **Anara for Word** in [Microsoft’s add-in marketplace](https://appsource.microsoft.com/en-us/marketplace/apps?search=anara\&product=office%3Bword), select **Get it now**, and follow the Microsoft account prompts to add it to Word. Open a document and launch Anara from **Home > Add-ins**. ## Research and write [#research-and-write] Ask Agent a specific question about the open document, or ask it to research a topic using your Anara library and academic sources. Review proposed edits in the document. In **Cite**, search for a paper and insert the reference where you need it. ```text title="Try it" Check the evidence behind this paragraph. Find sources for the claims and suggest a revision where the wording goes beyond the evidence. ``` ## If the button is missing [#if-the-button-is-missing] Open **Home > Add-ins** and select Anara from your installed add-ins. If your work or school account blocks the store, ask your Microsoft 365 administrator to allow Anara for Word. For account or usage problems, see [troubleshooting](/guides/troubleshooting). --- # Save pages with the Chrome extension Add web pages to your research library while you browse. The Anara Clipper saves web content to your library so you can search it, organize it, and ask questions about it later. ## Install and save [#install-and-save] 1. Open the [Anara Chrome extension](https://anara.com/chrome) and choose **Add to Chrome**. 2. Open Chrome's extensions menu and pin Anara if you want it visible in the toolbar. 3. Click the Anara icon and sign in. 4. Open the page you want to keep and use the extension to save it. 5. Find the saved page in your Anara library. Website items can be filtered by file type. ## Check the saved content [#check-the-saved-content] A saved link is only useful for research if Anara can retrieve its content. Pages behind sign-in screens or paywalls may not provide readable full text. Check the file's **Full text** status, then open it and confirm the text you need is available. For a PDF, download the file through your authorized access and [upload it](/guides/add-files). You can also [import a URL](/guides/import-by-url) without installing the extension. If saving fails, check your sign-in state in the extension, then retry. Include the page URL and error message when [contacting support](/guides/troubleshooting#contact-support). --- # How connectors work Connect reference managers, cloud storage, apps, and science databases. Connectors let Anara work with sources outside your library. Open **Settings > Connectors** to search the catalog, browse categories, and manage connected accounts. ## Imports and live access [#imports-and-live-access] An import copies selected content into your Anara library. That copy can be searched and cited after processing. It is not a continuously synchronized copy of the external file: import updated content when you need it. A live connector queries the connected service when the agent uses it. Adding a connection does not import everything in that account. The agent can search connected sources and, where supported, offer to import a result. ## Choose a connector [#choose-a-connector] | Connector | Plan access | Guide | | -------------------------------------- | ------------------- | ----------------------------------------------------------------------------------------------------- | | Zotero, Mendeley | Free and paid plans | [Zotero](/guides/zotero), [Mendeley](/guides/mendeley) | | Benchling, Synapse | Plus and above | Connect from the catalog and follow its provider-specific setup | | Public science databases | Plus and above | [Science databases](/guides/science-databases) | | Curated MCP apps, including Notion | Plus and above | [Apps and MCP servers](/guides/mcp-servers) | | Google Drive, Gmail, OneDrive, Dropbox | Pro and above | [Google Drive](/guides/google-drive), [Gmail](/guides/gmail), [OneDrive](/guides/onedrive-sharepoint) | | SharePoint | Enterprise and Max | [OneDrive and SharePoint](/guides/onedrive-sharepoint) | | X | Max | Search public posts through the connected account | | Custom MCP servers | Pro and above | [MCP servers](/guides/mcp-servers) | Legacy and negotiated plans can differ. The catalog shows the plan requirement for your workspace. Some connectors are unavailable in HIPAA-enabled workspaces. ## Connect and use a source [#connect-and-use-a-source] 1. Find the service in **Settings > Connectors** and click **Connect**. 2. Complete the provider's sign-in or setup instructions. 3. Return to Anara and check the connection status. 4. In chat, type `@` and select the source, or name the connected service in your request. Combine a connected source with your library when the question needs both. For example: "Find the project brief in Drive and compare it with the papers in my library." If an external action needs approval, review the proposed change before allowing it. Reconnect expired authorizations from Settings. For missing files, check the account, provider permissions, and [supported file types](/guides/file-types). --- # Connect Zotero Connect Zotero, import files, and use them in chat. Connect Zotero to find and import research without downloading each file first. This connector is available on every plan, including Free. ## Connect and import [#connect-and-import] 1. Open **Settings > Connectors**. 2. Find **Zotero** and click **Connect**. 3. Sign in to Zotero and authorize Anara. 4. Return to Anara and open the connected source's import picker. 5. Select the content you want and complete the import. 6. Check the imported files and their **Full text** status in your library. Select the library and collections containing your papers. A PDF linked only on your computer may not be available through Zotero online. Upload that PDF directly if it cannot be retrieved. ## Ask about your research [#ask-about-your-research] Type `@` and choose Zotero to direct the agent to the connected service. To work with copies already imported, open their folder or mention them in chat. ```text title="Prompt" Find the papers relevant to my project in Zotero. Show me the matches before importing them. Then compare the methods using the full text available in my library. ``` ## Keep content current [#keep-content-current] Imports are copies. Connecting an account does not guarantee every file is imported, and changes in Zotero do not automatically update those copies. Import new or changed content when needed and check the resulting records. If a file is missing, check its account, permissions, attachment, and format. Reconnect from Settings if authorization has expired. See [troubleshooting](/guides/troubleshooting) for processing failures. --- # Connect Mendeley Connect Mendeley, import files, and use them in chat. Connect Mendeley to find and import research without downloading each file first. This connector is available on every plan, including Free. ## Connect and import [#connect-and-import] 1. Open **Settings > Connectors**. 2. Find **Mendeley** and click **Connect**. 3. Sign in to Mendeley and authorize Anara. 4. Return to Anara and open the connected source's import picker. 5. Select the content you want and complete the import. 6. Check the imported files and their **Full text** status in your library. Choose the references and attached files you want to bring in. A reference record without its PDF cannot provide the full paper for passage-level answers. ## Ask about your research [#ask-about-your-research] Type `@` and choose Mendeley to direct the agent to the connected service. To work with copies already imported, open their folder or mention them in chat. ```text title="Prompt" Find the papers relevant to my project in Mendeley. Show me the matches before importing them. Then compare the methods using the full text available in my library. ``` ## Keep content current [#keep-content-current] Imports are copies. Connecting an account does not guarantee every file is imported, and changes in Mendeley do not automatically update those copies. Import new or changed content when needed and check the resulting records. If a file is missing, check its account, permissions, attachment, and format. Reconnect from Settings if authorization has expired. See [troubleshooting](/guides/troubleshooting) for processing failures. --- # Connect Google Drive Connect Google Drive, import files, and use them in chat. Connect Google Drive to find and import research without downloading each file first. This connector is available on Pro and above. ## Connect and import [#connect-and-import] 1. Open **Settings > Connectors**. 2. Find **Google Drive** and click **Connect**. 3. Sign in to Google Drive and authorize Anara. 4. Return to Anara and open the connected source's import picker. 5. Select the content you want and complete the import. 6. Check the imported files and their **Full text** status in your library. Use the Google account that can access the files you need. Browse the connected Drive and select the files or folders for import. Check the selected account if a shared file is missing. ## Ask about your research [#ask-about-your-research] Type `@` and choose Google Drive to direct the agent to the connected service. To work with copies already imported, open their folder or mention them in chat. ```text title="Prompt" Find the papers relevant to my project in Google Drive. Show me the matches before importing them. Then compare the methods using the full text available in my library. ``` ## Keep content current [#keep-content-current] Imports are copies. Connecting an account does not guarantee every file is imported, and changes in Google Drive do not automatically update those copies. Import new or changed content when needed and check the resulting records. If a file is missing, check its account, permissions, attachment, and format. Reconnect from Settings if authorization has expired. See [troubleshooting](/guides/troubleshooting) for processing failures. --- # Connect Gmail Find email context and bring relevant messages into your research. Connect Gmail to use email as a source in Anara. Gmail requires Pro or above. 1. Open **Settings > Connectors**. 2. Find **Gmail** and click **Connect**. 3. Sign in to the Google account you want and authorize access. 4. Return to Anara and check that the account is connected. ## Find email context [#find-email-context] Mention Gmail in chat and identify the sender, topic, or date range. Anara can search the connected account and use relevant messages in its response. ```text title="Prompt" Find the email about the revised study protocol from last week. Summarize the changes and compare them with the protocol in my project folder. ``` Gmail does not have the same file-import picker as Drive. Ask the agent to retrieve or import the relevant email content instead. Connecting Gmail does not import your entire mailbox. ## Check access [#check-access] If a message is missing, confirm the connected account and use a distinctive subject or sender. Reconnect from Settings when Google authorization expires. Review proposed external actions before approving them. See [connectors](/guides/connectors) for how live access differs from imported copies. --- # Connect OneDrive and SharePoint Import files from Microsoft storage and use them in research. OneDrive and SharePoint connect your Microsoft files to Anara. OneDrive requires Pro or above. SharePoint is available through Enterprise and Max entitlements. ## Connect [#connect] 1. Open **Settings > Connectors**. 2. Choose **OneDrive** or **SharePoint**, then **Connect**. 3. Sign in to the Microsoft account that can access the content. 4. Approve the requested access and return to Anara. 5. Open the source's import picker and select the files or folders you need. For SharePoint, choose the appropriate site and document library. Work or school accounts may require administrator consent. If Microsoft blocks authorization, ask your Microsoft 365 administrator to approve the connection. ## Use files in chat [#use-files-in-chat] Mention the connected source and name the file, folder, or site. Use your organization's own terminology to help locate it. For a sustained project, import the relevant files into an Anara folder and [chat with that folder](/guides/chat). ```text title="Prompt" Find the current project protocol in SharePoint and compare it with the studies in my project folder. Cite both sources and identify any missing evidence. ``` ## Missing or outdated content [#missing-or-outdated-content] Anara cannot access files the connected account lacks permission to read. Check the account, site, library, and file type. Imported files are copies; reimport changed content when needed. There is no automatic Resync step. See [connectors](/guides/connectors) for the distinction between imports and live access. --- # Connect Notion Search and update Notion through its hosted MCP connection. The Notion connector works with your live Notion workspace through MCP. It is available on Plus and above. It does not automatically copy your Notion workspace into your Anara library. ## Connect [#connect] 1. Open **Settings > Connectors**. 2. Find **Notion** and click **Connect**. 3. Complete Notion's authorization flow for the workspace you want to use. 4. Return to Anara and confirm the connection. ## Search alongside your library [#search-alongside-your-library] Type `@` and choose Notion, or ask Anara to find a named page or database in Notion. Include the page name or your team's terminology when possible. ```text title="Prompt" Find my Evidence Synthesis page in Notion. Compare its conclusions with the papers in my Anara project folder and cite the evidence for any disagreement. ``` For updates, describe the exact page and change you want. Review any approval request before allowing a write to Notion. The connected account's permissions still determine which content is accessible. ## Troubleshooting [#troubleshooting] Reconnect if authorization expires. If a page is missing, check that you connected the correct Notion workspace and granted access to that content. There is no Notion import connector or Resync workflow to configure. For other hosted apps, see [MCP servers](/guides/mcp-servers). --- # Search science databases Use public literature, trial, structure, and drug-discovery sources. Science connectors provide live access to public research databases. They require Plus or above and do not require a separate database login. | Source | Find | | ------------------- | --------------------------------------------------- | | PubMed | Biomedical literature records | | bioRxiv and medRxiv | Biology and health preprints | | ClinicalTrials.gov | Registered clinical studies | | Protein Data Bank | Protein and nucleic-acid structure records | | ChEMBL | Compounds, targets, assays, and bioactivity records | | Open Targets | Target-disease associations and supporting evidence | ## Connect and search [#connect-and-search] 1. Open **Settings > Connectors**. 2. Find the database and click **Connect**. 3. In chat, type `@` and select it, or name it in your question. 4. Include the identifiers, topic, dates, or filters that matter. ```text title="Prompt" Search ClinicalTrials.gov for studies matching my review criteria. Report their identifiers, status, and outcome measures, with links to the registry records. ``` Combine the database with your library to compare external records with files you already have. For paper discovery across disciplines, see [literature search](/guides/literature-search). ## Check what the result means [#check-what-the-result-means] A database record may provide metadata without full text. A preprint is not evidence of peer review, and a trial registration is not a published result. Open the cited record and confirm its status before using it in a conclusion. If results are incomplete, try the source's terminology or a specific identifier. --- # Apps and MCP servers Connect curated apps or add a custom MCP endpoint as an Anara source. Model Context Protocol (MCP) lets Anara use tools provided by another service. A connector can search or read that service and, when permitted, perform actions there. This page covers sources inside Anara. To connect another AI tool to your Anara library, use [Anara over MCP](/guides/mcp). ## Curated apps [#curated-apps] The catalog includes BioRender, Granola, Atlassian, Hugging Face, Notion, Airtable, PostHog, Zapier, Linear, Sentry, Intercom, Miro, ClickUp, Todoist, Stripe, and ElevenLabs. These curated connections require Plus or above. Provider subscriptions and permissions can impose additional requirements. 1. Open **Settings > Connectors** and find the app. 2. Click **Connect**. 3. Complete the provider's sign-in. If the dialog asks for a provider-specific server URL, use the instructions linked there. 4. Return to Anara and ask for a task using the connected app. The provider determines which tools are available. A connection does not mean every action in that product is supported. ## Add a custom server [#add-a-custom-server] Custom MCP servers require Pro or above. 1. In **Settings > Connectors**, choose **Add server**. 2. Paste the provider's HTTPS MCP endpoint. 3. Complete authentication when prompted. 4. Use **More options** only when the provider supplies a bearer token, custom headers, transport, or server name. 5. Confirm the connection and try a read request in chat. Review requests that would change external data before allowing them. Keep access tokens private. MCP connectors are disabled in HIPAA-enabled workspaces. ## Troubleshooting [#troubleshooting] Use the MCP endpoint, not the service's homepage. If authorization expires, reconnect. If no tools appear, verify that the endpoint exposes tools and that the connected account has permission to use them. --- # Anara over MCP Connect another AI assistant to your Anara workspace. Anara's remote MCP server lets another AI assistant work with your library and start research using Anara. It is separate from the [public docs MCP server](/guides/docs-for-ai) and from [MCP apps inside Anara](/guides/mcp-servers). ## Connect [#connect] Use a client that supports remote Streamable HTTP MCP with OAuth. Add a server with this URL: ```text https://anara.com/api/mcp ``` Complete the browser sign-in and review the consent screen. In clients that ask for a custom client ID or secret, leave those fields blank unless the client requires its own configuration. The exact place to add a server depends on your AI tool and its plan. ## Access and permissions [#access-and-permissions] The connection can only use workspace content your account is authorized to access. It can search and read sources and start Anara research. Read-only access is the default; optional write permission allows supported creation and editing actions. It does not grant unrestricted permission to delete, move, or rename library items. Review the consent scopes before connecting. Do not paste access tokens into chats or public configuration files. ## Available tools [#available-tools] | Tool | Purpose | | ------------------------ | ------------------------------------------------------------------------------- | | `anara_list_workspaces` | List accessible workspaces and identify the default | | `anara_execute` | Perform supported workspace operations under the granted permissions | | `anara_spawn_agent` | Start a research task and return a run identifier | | `anara_check_agent` | Check research status and incremental progress | | `anara_get_agent_result` | Read the completed result | | `anara_cancel_agent` | Request cancellation of a research run, then check status to confirm it stopped | The old `anara_run_code` tool is retired. Reconnect or refresh your client's tool list if it still offers old tool names. A research start returns a run identifier, not necessarily the finished answer; the client needs to check progress and retrieve the result. ## Troubleshooting [#troubleshooting] Confirm the endpoint above and complete OAuth sign-in. If access expires, authenticate again. If the wrong library appears, check the selected workspace. Anara research uses your account's applicable usage rules; see [usage and credits](/guides/usage-and-credits). For public paper search in your own application, see the [API reference](/api). --- # Workspaces and members Manage workspace identity, invitations, roles, and seats. A workspace holds a research library and its membership. Click the workspace name to switch between workspaces or create one. Check the selected workspace before changing billing or inviting collaborators. ## Workspace settings [#workspace-settings] Open **Settings > Workspace** to edit the name, workspace URL, or icon when you have permission. The URL slug is the part of the address that identifies the workspace. Update any saved links after changing it. ## Invite members [#invite-members] 1. Open **Settings > Members**. 2. Choose **Invite** and enter the email addresses. 3. Choose the role offered in the dialog. 4. Review any additional seats and charges, then send the invitation. Members can also join through the workspace invite link available on the Members page. Share it only with the people you intend to invite. If more paid seats are needed, the workspace owner must handle the additional seats. ## Roles [#roles] **Owners** manage workspace settings, billing, and membership. **Members** use the workspace and its shared content. The standard invitation dialog offers Owner and Member. Some account and billing operations also recognize an Admin role; that does not mean every workspace exposes Admin as an invitation option. Use the member's role control to change their access when permitted, or remove them from the workspace. Inviting someone to the workspace is different from [sharing one folder](/guides/collaborate). Use folder sharing when a collaborator only needs a particular project. ## Usage and billing [#usage-and-billing] Included usage is per person; purchased credits are shared by the workspace. Adding members can require paid seats. See [plans and billing](/guides/plans-and-billing) and [usage and credits](/guides/usage-and-credits). ## Leave or delete [#leave-or-delete] Eligible members can use **Settings > Workspace > Leave workspace**. Owners see **Delete workspace** instead. Deletion permanently removes the workspace and its associated data. Export anything needed and resolve membership and billing before confirming. If an ownership restriction blocks leaving, ask another owner or contact support. --- # Share folders and collaborate Give a collaborator access to a project and write together. Share a folder when someone needs a specific research project. Add them to a [workspace](/guides/workspaces) when they need to work with the team more broadly. ## Folder visibility [#folder-visibility] Private folders are restricted to their owner. Team folders are available to workspace members. Shared folders use explicit sharing controls. Check the folder's current visibility and member list before putting sensitive material in it. ## Share a folder [#share-a-folder] 1. Open the folder and its **Invite** or sharing control. 2. Choose the people or invite-link access you want to allow. 3. Select the role. 4. Share the invitation and review the folder's member list. | Role | Access | | ------------ | ---------------------------------------------------------------- | | Viewer | View files and members | | Collaborator | Add and remove files, invite members, and change folder settings | | Owner | Full control, including deleting the folder | Only folder owners can invite another owner. Roles and visibility affect which settings and actions are available. Folder sharing is also subject to your plan's sharing entitlement. ## Write together [#write-together] People with write access can edit shared notes collaboratively. Use [version history](/guides/version-history) to inspect earlier saved note versions. Keep project rules in [folder instructions](/guides/customize#folder-instructions). To remove access, update membership or visibility in the folder's settings and check the resulting member list. Workspace billing and membership are managed separately under [Settings > Members](/guides/workspaces). --- # Set your preferences Change language, appearance, model, citations, and notifications. Open **Settings > Preferences** to choose how Anara looks and how it responds. | Setting | What it changes | | ----------------------------- | ----------------------------------------------------------- | | Language | The interface and AI response language | | Theme | System, light, or dark appearance | | Enable auto-rename | Automatic file naming on import | | Include notes in AI responses | Whether AI can read and cite your notes | | Spellcheck | Spellchecking in the note editor | | AI model | Your default model for responses | | Citation format | The shared format across chats, notes, and exports | | Automatically rename chats | Whether chat titles can change as the conversation develops | | Notifications | Available notification preferences | Model choices depend on your plan. See [models and Deep Search](/guides/models-and-deep-search). For a journal-specific format or a custom CSL file, see [citation styles](/guides/citation-styles). ## Control response length [#control-response-length] Ask for the length in your message, such as "Give me a short answer" or "Explain this in detail." There is no separate response-length control in Preferences. ## Account and workspace settings [#account-and-workspace-settings] Use [Account and Devices](/guides/account-security) for your display name, two-step verification, and signed-in sessions. Use [Workspace and Members](/guides/workspaces) for the shared workspace. Subscription controls are in [Billing](/guides/plans-and-billing), and AI consumption is in [Usage](/guides/usage-and-credits). --- # Manage your account and security Update your name, enable two-step verification, and manage signed-in devices. Open **Settings > Account** to manage your identity and account access. Your display name is editable. Your email address is shown as read-only; contact support if you need help changing it. ## Enable two-step verification [#enable-two-step-verification] 1. In **Settings > Account > Security**, open **2-step Verification**. 2. Choose **Authenticator app (QR code)**. 3. Scan the QR code with your authenticator app. 4. Enter the six-digit code to verify setup. 5. Copy the backup codes and keep them somewhere you can access if you lose your device. 6. Complete the setup flow. Keep the backup codes private. If you lose both your authenticator and backup codes, [contact support](/guides/troubleshooting#contact-support). Turning two-step verification off requires an authenticator code. HIPAA-enabled workspaces do not allow it to be disabled. ## Devices [#devices] Open **Settings > Devices** to see signed-in sessions, the current device, and recent activity. Choose **Sign out** beside another device to revoke that session. **Sign out other devices** keeps the current session signed in and ends the others. **Settings > Account > Log out of all devices** ends every active session. **Clear cookies** removes local session data and requires you to sign in again; use it when troubleshooting a persistent session problem. ## Delete your account [#delete-your-account] Account deletion is permanent. [Export work](/guides/export) you want to keep and resolve any [workspace ownership](/guides/workspaces) first. Cancel subscriptions before deleting. App Store subscriptions must be cancelled through Apple and can keep billing after account deletion. Open **Settings > Account**, scroll to **Danger zone**, and choose **Delete**. Follow the confirmation flow. An active or past-due subscription you pay for can block deletion; use Billing or contact support to resolve it. For security information about Anara's service, see [security](https://anara.com/security). --- # Plans and billing Compare access, change a subscription, cancel, and find receipts. Your workspace's plan determines its model access, connectors, upload limits, and included usage. Open **Settings > Billing** to see the subscription for the selected workspace. Check [current pricing](https://anara.com/pricing) and the checkout total before buying. ## Choose a plan [#choose-a-plan] | Plan | Models and research modes | Connectors | | ------------------ | ------------------------------------------------------------------- | ------------------------------------------------------------- | | Free | Auto, daily usage allowance | Zotero and Mendeley | | Plus | Auto, larger weekly allowance | Free connectors, science databases, and curated apps | | Pro | Manual model selection, weekly allowance | Plus connectors, cloud storage, Gmail, and custom MCP servers | | Max 5x and Max 20x | Manual models, Deep Search, Model Council, larger weekly allowances | Pro connectors plus connectors available to Max | | Enterprise | Access depends on the Enterprise tier | Includes Enterprise connector access | Max 5x and Max 20x provide five and twenty times the standard Pro allowance respectively. They share the same research-mode entitlements. Existing or negotiated plans can differ, so use your workspace's Billing and Usage pages for its subscription and allowance. ## Upgrade or change billing frequency [#upgrade-or-change-billing-frequency] 1. Select the workspace you want to change. 2. Open **Settings > Billing** and choose **Upgrade** or the available plan action. 3. Select the plan and billing frequency. 4. Review the charge and effective date before confirming. The checkout or billing portal shows proration and the date a change takes effect. A subscription charge, a plan-change charge, and a credit purchase can appear as separate charges. ## Payment methods, invoices, and receipts [#payment-methods-invoices-and-receipts] Choose **Manage billing** beside **Billing details** to open payment methods, invoices, and receipts. If a payment fails, update the payment method there. If you cannot manage billing, ask the workspace owner to check it. ## Cancel [#cancel] For a web subscription, open **Settings > Billing > Cancel plan** and complete the cancellation flow. Check the displayed cancellation date. Cancelling stops future renewal at the end shown; it does not delete your account or library. For an Apple subscription, choose **Manage in the App Store** and cancel through Apple. Deleting the app or your Anara account does not cancel Apple billing. To move to web billing, first turn off Apple auto-renewal, then follow the option in Anara's Billing page and review when web billing begins. ## Unexpected charges and refunds [#unexpected-charges-and-refunds] Check the invoice description, workspace, date, and amount. [Credit purchases and auto-refills](/guides/usage-and-credits) are separate from your subscription. If a charge still looks wrong, open **Support > Contact us** and provide your account email, the receipt, date, amount, and reason for the request. The team can review the charge. A refund request does not cancel future subscription renewals. For App Store purchases, use Apple's billing and refund process. For referral or affiliate earnings, see [referrals and affiliates](/guides/referrals). --- # Usage and credits Understand your allowance, add credits, and recover from a usage limit. Anara measures the AI work you use rather than a fixed number of messages or words. Open **Usage** in the sidebar or **Settings > Usage** to see your meter, available credits, and the next reset time. ## What uses your allowance [#what-uses-your-allowance] Chat, AI editing, extraction in Views, and importing files use AI processing. Longer conversations, more source reading, heavier models, and Deep Search can use more than a short question. Reading saved files, organizing folders and tags, writing manually, and opening past answers do not require a new AI run. Free has a daily allowance; paid plans have a weekly allowance. The app shows the reset information for your account. Included usage is per person. Purchased credits belong to the workspace and can cover members' usage after their included allowance is used. ## When you reach a limit [#when-you-reach-a-limit] New AI requests and imports can be blocked while your saved work remains accessible. Read the notice: it tells you whether waiting for a reset will restore access or whether a credit balance needs attention. Do not assume that every block will clear at the next allowance reset. You can wait when a reset will restore access, change your [plan](/guides/plans-and-billing), or add credits on a paid plan. If the meter and error disagree, refresh and contact support with the task and error message. ## Add credits on the web [#add-credits-on-the-web] 1. Open **Settings > Usage** and choose **Add credits**. 2. Select an amount or enter your own. 3. Review the credit quantity and checkout total. 4. Complete checkout to charge the payment method and add the credits. Owners and admins can purchase credits. Members can use **Request credits** to send a request to the workspace's owners and admins. Free workspaces must upgrade before buying credits. Web credit purchases range from $5 to $1,000 in whole-dollar increments. There are 100 credits per US dollar. Unused credits bought on the web expire 12 months after purchase. Credit packs bought through Apple in the iOS app do not expire. Your included allowance is used before purchased credits. ## Auto-refill [#auto-refill] In the credit dialog, open **Auto-refill**. Set a balance threshold, refill amount, and monthly cap, then save. Review these values carefully because enabling auto-refill authorizes future charges when its conditions are met. You can change or disable it from Usage. If a purchase does not appear, refresh and check the receipt before trying another purchase. For payment failures or unexpected charges, see [billing](/guides/plans-and-billing). --- # Referrals and affiliates Find your referral link and review partner-program terms. Use **Invite and earn** in Anara's settings navigation to open the partner portal and get your referral link. The portal manages your link, rewards, and payout details. ## Share a referral link [#share-a-referral-link] 1. Open **Invite and earn** while signed in to Anara. 2. Sign in to the partner portal if prompted. 3. Review your program and payout details. 4. Copy your link and share it with people who may find Anara useful. Check the current terms in the portal before promising a discount or payout. Referral earnings are separate from the credits used for AI work in Anara. ## Affiliate program [#affiliate-program] Content creators can review the affiliate program at the [Anara partner portal](https://partners.dub.co/anara). Check its current commission, attribution period, payout method, and promotion rules before joining or running a campaign. If a referral is missing or a payment looks wrong, contact support with your partner account and the relevant referral or payout details. Do not share another person's payment information in a public message. --- # Troubleshooting and support Resolve upload, processing, chat, connector, sign-in, and billing problems. Start with the exact error and the item it affects. A file problem, expired connection, and usage limit need different fixes. ## Uploads and processing [#uploads-and-processing] * Check the [supported format, file size, and page limit](/guides/file-types). * Keep the connection active while uploading. Refresh to check status before starting another upload. * Check **Usage**. Imports need available processing allowance or credits. * For a damaged file, download or export a fresh copy from its source application. * For a password-protected file, upload an unlocked copy you are authorized to use. * If Word or PowerPoint content extracts poorly, try a PDF export from the original application. **Abstract only** usually means the full paper was not available from the imported link. Upload a full PDF through your authorized access. **Content unavailable** means readable content was not obtained. Check the failure message before retrying. Scanned PDFs need OCR. Anara's OCR fallback is available on Pro and above. If your plan or the file prevents OCR, upload a searchable PDF with a text layer. A processing failure does not always mean the file is empty. ## Missing or incorrect answers [#missing-or-incorrect-answers] Check the source controls and the file's full-text status. Ask the question with a file or folder mention, then open the citations and compare them with the answer. For a wrong claim, identify the passage and explain the discrepancy in a follow-up. Use the response feedback controls to report the problem. If an answer stops early, ask it to continue. For a failed response, retry or choose another available model. Persistent errors belong in a support report rather than repeated identical requests. ## Connectors [#connectors] Open **Settings > Connectors** and check the account and status. Reconnect expired authorizations. For work or school Microsoft accounts, administrator consent may be needed. Imported copies do not update automatically when the source changes; import the updated content. ## Sign-in and devices [#sign-in-and-devices] Use the same account and sign-in method you used before. If already signed in, **Settings > Account > Clear cookies** can reset local session state; it signs you out. For a lost authenticator, use your saved backup codes or contact support. See [account security](/guides/account-security). ## Billing and usage [#billing-and-usage] Check the selected workspace, then **Settings > Billing** and **Settings > Usage**. A credit top-up is a separate charge from a subscription. Apple subscriptions are managed through Apple. See [plans and billing](/guides/plans-and-billing) and [usage and credits](/guides/usage-and-credits). ## Contact support [#contact-support] Open **Support** at the bottom of the sidebar and choose **Contact us** for account, billing, or technical help. Use **Send feedback** for a bug report or product suggestion. Include your account email, the task you attempted, the exact error, the file or connector involved, and steps to reproduce it. For an unexpected charge, include the date, amount, and receipt. Never include your password, backup codes, or access tokens. --- # Use these docs with AI Give your coding agent or AI assistant the Anara docs as clean Markdown, a single file, or an MCP server. Every page on this site is written to be read by agents as well as people. Pick whichever fits your tool. ## Copy a page [#copy-a-page] Use **Copy for LLM** in the right column of any page to copy it as Markdown, ready to paste into a chat. **View as Markdown** opens the same text in the browser. ## Fetch pages as Markdown [#fetch-pages-as-markdown] Append `.md` to any docs URL, or send an `Accept: text/markdown` header to the normal URL. ```bash title="Terminal" curl https://docs.anara.com/guides/quickstart.md curl -H "Accept: text/markdown" https://docs.anara.com/api/search ``` ## Index and full text [#index-and-full-text] | File | What it holds | | ---------------------------------- | ------------------------------------------------------------ | | [`/llms.txt`](/llms.txt) | Every page with a one-line summary, grouped like the sidebar | | [`/llms-full.txt`](/llms-full.txt) | The full text of every page in one file | | [`/sitemap.xml`](/sitemap.xml) | Every page URL | ## Connect the docs over MCP [#connect-the-docs-over-mcp] The docs are also a public MCP server (Streamable HTTP, no sign-in) at `https://docs.anara.com/mcp`. It has three tools: * `list_docs`: every page with its URL and summary * `search_docs`: full-text search across all pages * `read_doc`: one page as Markdown ```json title="MCP client config" { "mcpServers": { "anara-docs": { "url": "https://docs.anara.com/mcp" } } } ``` This server only reads the docs. To let an agent work with your own Anara library, see [Anara over MCP](/guides/mcp). --- # Getting started Find papers and search the literature with one API key. The Anara API provides paper lookup and academic search. Search in instant, standard, or deep mode. Full-text passages and hosted PDFs are available for papers in our index unless the paper is paywalled or its licence forbids commercial reuse; those papers return metadata and a short abstract excerpt. Try the [playground](https://platform.anara.com/playground) with your Anara sign-in. An API key is only needed when calling the API from your own application. Use the API URL shown in **Get code** in your console. Preview keys belong to the preview environment and do not authenticate against the production API. Add prepaid credit in the console to use paid endpoints. See [Billing](/api/billing). ## 1. Get a key [#1-get-a-key] Create a key in the console under [API keys](https://platform.anara.com/keys). Keys start with `ana_live_`. The full key is shown once, at creation. After that the console shows only the prefix. Send the key as a bearer token or in the `X-API-Key` header. ```http title="Header" Authorization: Bearer $ANARA_API_KEY ``` A missing, unknown, or revoked key answers `401`. A key without the scope for the product answers `403`. See [Keys and scopes](/api/keys). ## 2. Make a request [#2-make-a-request] ```bash title="cURL" curl -X POST https://api.anara.com/v1/search \ -H "Authorization: Bearer $ANARA_API_KEY" \ -H "Content-Type: application/json" \ -d '{"query": "transformer attention efficiency", "maxResults": 5}' ``` ```python title="Python" import os import requests response = requests.post( "https://api.anara.com/v1/search", headers={"Authorization": f"Bearer {os.environ['ANARA_API_KEY']}"}, json={"query": "transformer attention efficiency", "maxResults": 5}, ) print(response.json()) ``` ```ts title="TypeScript" const response = await fetch('https://api.anara.com/v1/search', { method: 'POST', headers: { Authorization: `Bearer ${process.env.ANARA_API_KEY}`, 'Content-Type': 'application/json', }, body: JSON.stringify({ query: 'transformer attention efficiency', maxResults: 5, }), }); console.log(await response.json()); ``` ```json title="Response" { "papers": [ { "id": "W2963403868", "title": "Attention Is All You Need", "authors": ["Ashish Vaswani", "Noam Shazeer"], "year": 2017, "doi": "10.48550/arxiv.1706.03762", "citedByCount": 98000, "openAccess": true, "hasFulltext": true } ], "totalApprox": 1, "charge": { "costMicroUsd": 5000, "balanceMicroUsd": 9995000 } } ``` Every metered response carries a `charge` object: what the call cost and what is left, both in micro-dollars (1 USD = 1,000,000). ## 3. Read the reference [#3-read-the-reference] ## Copy documentation [#copy-documentation] Use **Copy page as text**, [llms.txt](/llms.txt), or [llms-full.txt](/llms-full.txt). Each documentation page is also available at `/.md`. --- # Use with coding agents Set up Claude Code, Cursor, Codex, or any coding agent to call the Anara API correctly on the first try. Coding agents do best when they have a key in the environment, a short note on how the API works, and docs they can read as Markdown. This page gives you all three. ## 1. Put your key in the environment [#1-put-your-key-in-the-environment] Create a key under [API keys](https://platform.anara.com/keys) in the console, then export it. Agents should read it from the environment and never paste it into code. ```bash title="Terminal" export ANARA_API_KEY="ana_live_..." ``` ## 2. Add the API to your agent instructions [#2-add-the-api-to-your-agent-instructions] Paste this into `AGENTS.md` at the root of your repo (Codex, Cursor, and most agents read it; Claude Code reads `CLAUDE.md`, so put it there too, or import `AGENTS.md` from it). ```md title="AGENTS.md" ## Anara API (academic paper search) - Base URL: https://api.anara.com. Auth: `Authorization: Bearer $ANARA_API_KEY`. Never hardcode the key. - Search papers: `POST /v1/search` with `{"query": "...", "type": "standard", "maxResults": 10}`. `type` is `instant` (fastest), `standard` (default, hybrid), or `deep` (full text with citable passages where the licence allows). Optional `filters`, for example `{"publicationYear": ">=2021", "hasFulltext": true}`. - Resolve known papers: `POST /v1/resolve` with `{"identifiers": [...]}` (DOI, arXiv, OpenAlex, or PubMed ids) and/or `{"titles": [...]}`. - Application errors are JSON with a `_tag` and a `message`; schema-invalid requests can return `400` with an empty body. `402 PaymentRequired` means the workspace needs credit. - Docs as Markdown: https://docs.anara.com/api/llms.txt lists every API page; append `.md` to any docs URL. - OpenAPI spec: https://api.anara.com/openapi.json ``` ## 3. Let the agent read the docs [#3-let-the-agent-read-the-docs] * [`/api/llms.txt`](/api/llms.txt) lists only the API pages, each as a Markdown link. * Append `.md` to any page URL, for example `https://docs.anara.com/api/search.md`. * The docs MCP server at `https://docs.anara.com/mcp` lets an agent search and read every page. See [Use these docs with AI](/guides/docs-for-ai) for the client config. ## 4. Generate a typed client (optional) [#4-generate-a-typed-client-optional] If you want types instead of raw requests, generate a client from the spec. See [OpenAPI and SDKs](/api/openapi). ## Check it works [#check-it-works] Ask your agent to run a small search and print the titles. If it gets `401`, the key is missing or revoked; if it gets `403`, inspect the error tag for scope, platform-access, or HIPAA restrictions. See [Statuses](/api/statuses) for everything else. --- # Search Search the academic index with filters. `POST /v1/search` runs a metadata search over the corpus: titles, abstracts, authors, venues. It needs the `papers` scope. Instant costs **$2**, standard **$5**, and deep **$12 per 1,000 queries**. The number of results does not change the price. ## Request [#request] | field | type | notes | | ----------------- | ---------------------------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | `query` | `string` | Natural language or keywords. 1 to 1,000 characters. | | `type` | `"instant" \| "standard" \| "deep"` | `instant` is vector-only and fastest. `standard` (default) is hybrid keyword plus vector. `deep` also searches full text and returns citable passages where we can share them. | | `maxResults` | `int` | Default 10, max 100. | | `filters` | object | See below. | | `sort` | `"relevance" \| "recent" \| "citations"` | `relevance` (default) ranks by fit to the query, breaking near-ties toward papers cited unusually often for their age. `recent` returns the most relevant papers from the last two years, widening to four when few match, and demotes repository deposits. `citations` orders the relevant papers by citation count. Works with every `type`. | | `includeTotal` | `boolean` | Best effort: include `totalEstimate` only if it is ready when results arrive. Prefer `/v1/search/total`. | | `includePassages` | `boolean` | **Admin only.** Attach up to two full-text passages per paper on `instant` and `standard` searches, where we can share them; `deep` always carries passages. Requests from non-admin keys are rejected with `400`. | ### Filters [#filters] | field | type | example | | ----------------- | --------- | ----------------------------------------------- | | `publicationYear` | `string` | `"2020"`, `">=2018"`, `"2015..2020"` | | `isOpenAccess` | `boolean` | `true` | | `hasFulltext` | `boolean` | `true` restricts to papers that `deep` can read | | `field` | `string` | `"Computer Science"` | | `citedByCount` | `string` | `">100"` | ```bash title="cURL" curl -X POST https://api.anara.com/v1/search \ -H "Authorization: Bearer $ANARA_API_KEY" \ -H "Content-Type: application/json" \ -d '{"query": "CRISPR off-target detection", "maxResults": 20, "filters": {"publicationYear": ">=2021", "hasFulltext": true}}' ``` ```python title="Python" import os import requests response = requests.post( "https://api.anara.com/v1/search", headers={"Authorization": f"Bearer {os.environ['ANARA_API_KEY']}"}, json={"query": "CRISPR off-target detection", "maxResults": 20, "filters": {"publicationYear": ">=2021", "hasFulltext": True}}, ) print(response.json()) ``` ```ts title="TypeScript" const response = await fetch('https://api.anara.com/v1/search', { method: 'POST', headers: { Authorization: `Bearer ${process.env.ANARA_API_KEY}`, 'Content-Type': 'application/json', }, body: JSON.stringify({ query: 'CRISPR off-target detection', maxResults: 20, filters: { publicationYear: '>=2021', hasFulltext: true }, }), }); console.log(await response.json()); ``` ## Response [#response] ```json title="Response" { "papers": [ { "id": "W4210000000", "title": "…", "year": 2022, "authors": ["…"], "abstract": "…", "citedByCount": 41, "openAccess": true, "hasFulltext": true } ], "totalApprox": 20, "charge": { "costMicroUsd": 5000, "balanceMicroUsd": 9990000 } } ``` `totalApprox` is the number of papers in this response. It is not a count of the whole corpus. `totalEstimate` appears only when the request sets `includeTotal` and the estimate is ready by the time results are. Search never waits for it. To always get the count, call `/v1/search/total` alongside the search. ## Total matches [#total-matches] `POST /v1/search/total` estimates how many papers in the corpus have every word of the query in their title or abstract, under the given filters, rounded to two significant figures, like the hit count a search engine shows. Counts under 1,000 are exact. Send it in parallel with `/v1/search`, with the same `query`, `type` and `filters`, so results never wait for the count. It is free and needs the `papers` scope. | field | type | notes | | --------- | ------------------------- | -------------------------------------- | | `query` | `string` | The same query as the search. | | `type` | `"instant" \| "standard"` | Default `standard`. Deep has no count. | | `filters` | object | The same filters as the search. | The response is `{ "totalEstimate": 19000 }`. `totalEstimate` is `null` when the query has no words to count, for example only stopwords. It usually arrives in under a second. ## Deep search [#deep-search] Use `POST /v1/search` with `"type": "deep"`, or select **Deep** under **Type** in the playground. Deep search also searches full text and returns citable passages where we can share them; other papers come back with metadata and an abstract excerpt. Add `isOpenAccess: true` to restrict results to open access. Deep returns `totalApprox` without a corpus estimate. ## Result count and coverage [#result-count-and-coverage] The response contains up to `maxResults` ranked papers. Call `/v1/search/total` for the estimated match count. Passages, hosted PDF links and full abstracts are returned unless the paper is paywalled or its licence forbids commercial reuse (CC BY-NC variants). For those papers, `abstract` is a short excerpt matched to your query and `abstractIsExcerpt` is `true`. A missing PDF link means no hosted copy is available; the API does not fall back to a publisher download. Additional filters include `authorName`, `sourceName`, `field`, `citedByCount`, `publicationYear`, `isOpenAccess`, and `hasFulltext`. `hasFulltext` returns only papers with full text we can serve. Workspace requests are limited to 120 per minute across API keys. The signed-in playground has a separate limit of 30 per minute. A `429` response means retry after 60 seconds. Each paper includes `venue`, `publicationDate` (the full indexed date, not just `year`), `field`, `language`, `arxivId`, `pmcid`, and `pmid` alongside `doi` when available. Missing metadata is omitted. `openAlexId` is not exposed. --- # Resolve Resolve DOIs, arXiv, OpenAlex, PubMed ids, or exact titles to papers. `POST /v1/resolve` resolves up to 100 identifiers and 50 titles in one call. It needs the `papers` scope. Lookups are **free up to 5,000 per calendar month** per workspace. After that they cost $0.10 per 1,000. One call is one lookup, whatever it contains. ## Request [#request] | field | type | notes | | ------------- | ---------- | ----------------------------------------------------------------------------------------------------------------------------- | | `identifiers` | `string[]` | DOI (`10.1038/...`), arXiv (`arXiv:2106.09685` or `2106.09685`), OpenAlex (`W2963403868`), PubMed (`pmid:31234567`). Max 100. | | `titles` | `string[]` | Exact or near-exact titles. Max 50. | | `maxResults` | `int` | Cap on papers returned. Default 10, max 100. | Send at least one of `identifiers` or `titles`. ```bash title="cURL" curl -X POST https://api.anara.com/v1/resolve \ -H "Authorization: Bearer $ANARA_API_KEY" \ -H "Content-Type: application/json" \ -d '{"identifiers": ["10.1038/s41586-020-2649-2", "arXiv:1706.03762"]}' ``` ```python title="Python" import os import requests response = requests.post( "https://api.anara.com/v1/resolve", headers={"Authorization": f"Bearer {os.environ['ANARA_API_KEY']}"}, json={"identifiers": ["10.1038/s41586-020-2649-2", "arXiv:1706.03762"]}, ) print(response.json()) ``` ```ts title="TypeScript" const response = await fetch('https://api.anara.com/v1/resolve', { method: 'POST', headers: { Authorization: `Bearer ${process.env.ANARA_API_KEY}`, 'Content-Type': 'application/json', }, body: JSON.stringify({ identifiers: ['10.1038/s41586-020-2649-2', 'arXiv:1706.03762'], }), }); console.log(await response.json()); ``` ## Response [#response] ```json title="Response" { "papers": [ { "id": "W3103145119", "title": "Array programming with NumPy", "year": 2020, "doi": "10.1038/s41586-020-2649-2", "authors": ["Charles R. Harris"], "citedByCount": 12000, "openAccess": true, "hasFulltext": true }, { "id": "W2963403868", "title": "Attention Is All You Need", "authors": ["Ashish Vaswani"], "year": 2017, "arxivId": "1706.03762", "citedByCount": 100000, "openAccess": true, "hasFulltext": true } ], "resolutions": [ { "input": "10.1038/s41586-020-2649-2", "status": "resolved", "paperIds": ["W3103145119"] }, { "input": "arXiv:1706.03762", "status": "resolved", "paperIds": ["W2963403868"] } ], "charge": { "costMicroUsd": 0, "balanceMicroUsd": 10000000 } } ``` `resolutions` has one entry per input. `status` is `resolved`, `ambiguous` (a title matched more than one paper; all ids are listed), or `unresolved`. Each paper includes `venue`, `publicationDate` (the full indexed date, not just `year`), `field`, `language`, `arxivId`, `pmcid`, and `pmid` alongside `doi` when available. Missing metadata is omitted. `openAlexId` is not exposed. --- # Billing Prepaid credit, the rate card, and top-ups. The API is prepaid. Each workspace has a credit balance. Metered calls debit it before the work runs. If the balance cannot cover a paid call, the API answers `402` until you top up. The first 5,000 resolve calls per workspace each calendar month are free and do not require a positive balance. There is no automatic monthly cash-credit grant. ## Rate card [#rate-card] | operation | price | notes | | ----------------- | ---------------------------------------- | -------------------------------------- | | Resolve | free to 5,000 per month, then $0.10 / 1k | one call is one lookup | | Search (instant) | $2 / 1k queries | result count does not change the price | | Search (standard) | $5 / 1k queries | result count does not change the price | | Search (deep) | $12 / 1k queries | beta price, may change | Prices are per query, not per result. `maxResults` defaults to 10 and caps at 100. ## Reading charges [#reading-charges] Every metered response carries `charge`: ```json title="Response" "charge": { "costMicroUsd": 5000, "balanceMicroUsd": 9995000 } ``` Values are in micro-dollars: divide by 1,000,000 for USD. `GET /v1/account/balance` returns the same balance with the open credit lots and the rate card. ## Top-ups [#top-ups] Buy credit in the console under **Billing**. Top-ups are $10 to $50,000, paid by card through Stripe, and earn a bonus at volume: | top-up | bonus | | ------------ | ----- | | from $1,000 | 10% | | from $5,000 | 20% | | from $20,000 | 30% | Purchased credit is valid for 12 months. Free credit is spent before purchased credit. ## Auto top-up [#auto-top-up] Turn on auto top-up in **Billing** to charge your saved card when the balance drops below a threshold you pick. The API checks after each debit and tops up at most once every ten minutes. --- # Keys and scopes Create, scope, inspect, and revoke API keys. Keys belong to a workspace. Workspace owners and admins create and revoke them in the console. Members can see the prefix, name, and usage of each key. ## Scopes [#scopes] Each key carries one or both scopes: | scope | grants | | -------- | ----------------------------------------- | | `papers` | `/v1/search` and `/v1/resolve` | | `pdf` | No public endpoint uses this scope today. | Calling an endpoint outside the key's scopes answers `403 InsufficientScope`. Pick the smallest set the integration needs. ## Inspect the current key [#inspect-the-current-key] ```bash title="cURL" curl https://api.anara.com/v1/keys/current \ -H "Authorization: Bearer $ANARA_API_KEY" ``` ```python title="Python" import os import requests response = requests.get( "https://api.anara.com/v1/keys/current", headers={"Authorization": f"Bearer {os.environ['ANARA_API_KEY']}"}, ) print(response.json()) ``` ```ts title="TypeScript" const response = await fetch("https://api.anara.com/v1/keys/current", { method: "GET", headers: { Authorization: `Bearer ${process.env.ANARA_API_KEY}`, }, }); console.log(await response.json()); ``` ```json title="Response" { "id": "k_…", "name": "prod", "keyPrefix": "ana_live_Ab12", "scopes": ["papers"], "workspaceId": "…", "createdAt": "…", "revokedAt": null } ``` ## Revoke [#revoke] Revoke in the console, or let the key revoke itself, for example from an incident runbook: ```bash title="cURL" curl -X POST https://api.anara.com/v1/keys/current/revoke \ -H "Authorization: Bearer $ANARA_API_KEY" ``` ```python title="Python" import os import requests response = requests.post( "https://api.anara.com/v1/keys/current/revoke", headers={"Authorization": f"Bearer {os.environ['ANARA_API_KEY']}"}, ) print(response.json()) ``` ```ts title="TypeScript" const response = await fetch("https://api.anara.com/v1/keys/current/revoke", { method: "POST", headers: { Authorization: `Bearer ${process.env.ANARA_API_KEY}`, }, }); console.log(await response.json()); ``` Revocation takes effect at once. Every later request with that key answers `401` with the message `API key revoked`. --- # Statuses Every status the API returns and what to do about it. Application errors are JSON with a `_tag` and a `message`. Request-schema validation failures return `400` and may have an empty body, so check the HTTP status before parsing JSON. | status | `_tag` | meaning | what to do | | ------ | --------------------------------- | ----------------------------------------------------------------------------- | --------------------------------------------------------------------- | | `400` | `InvalidRequest` or an empty body | The request is invalid. | Check required fields, types, and limits against the request schema. | | `401` | `Unauthorized` | Key missing, unknown, or revoked. | Send a live key as `Authorization: Bearer` or `X-API-Key`. | | `402` | `PaymentRequired` | No credit left for this call. `topUpUrl` points at the console. | Top up in the console. | | `403` | `InsufficientScope` | The key lacks the scope for this product. | Create a key with the `papers` scope. | | `429` | `RateLimited` | The workspace exceeded its request limit. | Wait 60 seconds before retrying. | | `403` | `HipaaRestricted` | The workspace is HIPAA-enabled; Papers is not available. | Use a non-HIPAA workspace. | | `502` | `UpstreamUnavailable` | The search index did not answer. | Retry with backoff. Metered calls that fail this way are not charged. | | `503` | `ContentsUnavailable` | Public contents retrieval is intentionally disabled pending licensing review. | Do not retry; no retrieval or charge occurs. | ## Retrying [#retrying] Treat `502` as retryable after a short backoff and `429` as retryable after 60 seconds. Treat `402` as final until you top up. --- # OpenAPI and SDKs Generate a client from the spec. The API is described by an OpenAPI 3.1 document generated from the same types that validate every request and response. * Spec: `https://api.anara.com/openapi.json` * Interactive reference: `https://api.anara.com/docs` Application errors are tagged objects with a `_tag` and a `message`. Request-schema validation failures return `400` and may have an empty body; check the status before parsing JSON. Errors that need more carry it as fields, for example `PaymentRequired` has `balanceMicroUsd`, `requiredMicroUsd`, and `topUpUrl`. The spec lists each endpoint's possible errors by status, so a generated client can narrow on them. ## Generate a client [#generate-a-client] ```bash title="Terminal" npx @hey-api/openapi-ts -i https://api.anara.com/openapi.json -o src/anara -c @hey-api/client-fetch ``` ```bash title="Terminal" openapi-generator-cli generate -i https://api.anara.com/openapi.json -g python -o anara-python ``` Send the key as a bearer token; both `bearer` and `apiKey` security schemes are in the spec. ## For agents [#for-agents] The docs are readable by machines as well as people. * `https://docs.anara.com/api/llms.txt` lists the API pages with a summary. `https://docs.anara.com/llms.txt` covers the whole site. * [Use with coding agents](/api/coding-agents) has a ready `AGENTS.md` snippet. * `https://docs.anara.com/llms-full.txt` is the full text of every page in one file. * Append `.md` to any docs URL, or send `Accept: text/markdown`, to get one page as Markdown. * `https://docs.anara.com/mcp` is an MCP server (Streamable HTTP, no auth) with `list_docs`, `search_docs`, and `read_doc` tools. ```json title="MCP client config" { "mcpServers": { "anara-docs": { "url": "https://docs.anara.com/mcp" } } } ``` ## API version [#api-version] The public endpoints are `/v1/search` with `type` (`instant`, `standard`, or `deep`) and `/v1/resolve`. This pre-launch API replaces the earlier endpoint names without compatibility aliases. Once integrations are in use, incompatible request/response changes require a versioned migration.