Skip to main content
Paste a fixed extraction template and attach a paper. Anara reads the document and fills one row of your table with the values it finds. Every cell traces to a specific passage. Customize the template for your field. Run the same extraction against every paper in a folder using Deep Search. For building the research log you return to throughout your project.

​
1. Describe the task

You have forty papers to read and a research log to fill. Reading each paper end to end to extract the same eight fields is the kind of work that takes a day per batch and drifts in consistency as the project stretches into months. Anara can run a fixed extraction template against each paper and return a filled row. You define the fields. You attach the paper. Anara reads it and returns only what the paper actually reports. Cells that cannot be filled from the text stay blank rather than filled with something plausible-sounding. The same template works across fields. A materials science researcher and a public health researcher run templates that look nearly identical. Their fields are different, but the motion is the same: one paper in, one structured row out, consistent across the entire project. Here a PhD student in biomedical engineering has twenty-three papers on hydrogel biomaterials for a graduate course project and needs a summary log she can sort and compare. 
Please create a summary table for this research paper using the following fields:

Title | Authors | Year | Research question or objective | Study type and methodology | Key findings | Limitations | Conclusions and implications

Fill each cell using only what this paper reports. If a field is not addressed in the paper, write "Not reported." Quote the specific passage where the key finding appears. Do not add interpretation beyond what the paper states.

​
2. Give Anara context

Required context
  • The paper, either uploaded to your library or attached to the chat.
Optional context
  • A custom field list if the default template does not match your research area. Anara replaces the default fields with whatever you specify.
  • One completed example row. Helps Anara match your level of detail and your phrasing conventions.

​
3. What Anara creates

A filled table row, formatted as a markdown table, with each cell citing the passage it came from. The output is your research log entry for this paper. You copy it into your notes or review tracker and open a new chat for the next paper. Over a full literature review, these rows become the working dataset you compare, filter, and sort.

​
4. Follow-up prompts

​
Add a field Anara missed

When the default extraction skipped something you need. 
Add a column for "Sample size" between Methodology and Key findings, and fill it for this paper. If sample size is not reported separately, check the methods section for participant count.

​
Run the batch across a folder

When you are ready to process a full set of papers at once. Requires Deep Search. 
Run this same extraction template across every paper in my folder "Hydrogel Biomaterials." Return one table with all rows combined. Use Deep Search so you can process the papers in parallel. Keep the same column structure.

​
Catch a cell that looks too full

When a cell contains more detail than the paper seems to support. 
For the Key Findings cell you filled: quote the exact passage from the paper where this is stated. If you summarized across multiple passages rather than quoting one, show me the passages you drew from.

​
5. Tricks, tips, and troubleshooting

​
How you word your prompt shapes what you get

The instruction “use only what this paper reports” is the most important line in the prompt. Without it, Anara completes sparse cells by drawing on what it knows about the field, which introduces errors that are hard to catch later. Asking for a quoted passage in the key findings cell turns every row into a self-auditing unit: if the passage is not there, the finding is not supported. If your extraction template has more than ten fields, split it into two passes: a descriptive pass (study type, population, methods) and an outcomes pass (findings, limitations, implications). Long templates produce compressed cells.

​
Check the output against your own understanding

Read the quoted passage before you copy the row. A cell that looks well-filled but does not appear in the quoted passage is the failure mode to catch. This happens most often in the methodology cell, where Anara may describe the study type using a label the paper does not use itself. When you disagree with a classification, ask Anara to show you the specific sentence in the methods section it based the entry on. If the paper uses a non-standard study design, add a note to that cell in your own words.

​
What to do with the output next

Paste the row into your tracking spreadsheet or note alongside the other extracted rows. Over time, sorting by the methodology column or the limitations column surfaces patterns across the literature that are hard to see paper by paper. When the full set is extracted, use the compare-papers workflow to probe specific dimensions across a subset, or ask Anara to write a synthesis paragraph from the rows you have collected.