Skip to main content
Define your extraction schema once, attach a paper, and Anara fills every field using only what the paper reports. Study type, sample size, intervention, outcomes, limitations: all populated from the source, with passages quoted so you can verify. For systematic reviewers working through a bibliography one paper at a time.

​
1. Describe the task

You have an inclusion list and a schema. What you do not have is time. Working through forty papers by hand, field by field, is where systematic reviews stall: not in the search or the screening, but in the extraction. Anara can take one paper and a defined schema and fill every extractable field from what the paper actually reports. Any field the paper does not address stays blank. Nothing is inferred from general clinical knowledge. The output quotes the passages it used, so you can check each entry before it goes into your review log. Here a PhD student in public health has 62 included papers for a systematic review on physical activity interventions in adolescents, a fixed eight-field extraction template, and is working through one paper per session. 
I am conducting a systematic review on physical activity interventions in adolescents aged 10-18. I have a fixed extraction schema. For the attached paper, fill in every field below using only what the paper reports. If a field is not addressed, write "Not reported." Do not infer or generalize from other studies.

Fields to extract:
- Study design
- Sample size and population
- Country
- Intervention type and duration
- Control condition
- Primary outcome measure and result
- Secondary outcomes
- Limitations stated by the authors

At the bottom, quote the three passages from the paper you relied on most heavily for the outcomes fields.

​
2. Give Anara context

Required context
  • The paper as a PDF attachment or imported to your library.
  • Your extraction schema: the field names and any specific definitions you use (for example, what counts as a “physical activity intervention” in your protocol).
Optional context
  • One or two rows you have already completed. Anara matches the level of detail and field format you established.
  • Your inclusion/exclusion criteria. Helps Anara flag anything that might affect the paper’s applicability before you extract from it.

​
3. What Anara creates

A completed extraction table row, one field per line, drawn strictly from the attached paper. Below the row, Anara quotes the specific passages it used for the outcome fields, so you can verify each data point before it goes into your review. Fields the paper does not address are labeled “Not reported” to keep the log consistent across papers. The row is the unit you paste into your extraction spreadsheet and move on from.

​
4. Follow-up prompts

​
Screen an abstract before you commit to importing

When you want to apply inclusion and exclusion criteria to an abstract before pulling the full paper. 
Here is an abstract from a paper I am deciding whether to include. Apply my criteria below and tell me: include, exclude, or uncertain. Give a one-sentence rationale for each criterion that applies.

Inclusion criteria: RCT or quasi-experimental design, adolescent population (10-18), physical activity as primary intervention, published 2010 or later.
Exclusion criteria: clinical population only, no control group, conference abstract.

Abstract: [paste abstract text here]

​
Run the same extraction across a full folder

When you are ready to move from one paper at a time to batch processing across a folder. 
I have a folder called "Included Studies" with 20 papers ready for extraction. Use Deep Search to run the same extraction schema across every paper in this folder and return one completed row per paper. Flag any paper where a field could not be reliably extracted. Use only what each paper reports.

Extraction fields: [paste your schema here]

​
Compare two papers on a contested outcome

When two included papers report the same outcome differently and you want to understand the divergence. 
Two papers in my extraction log report physical activity frequency as a primary outcome but arrive at different effect estimates. Pull the relevant passages from both papers and tell me whether the difference is likely explained by measurement approach, population, or something else the papers name explicitly.

​
5. Tricks, tips, and troubleshooting

​
How you word your prompt shapes what you get

The phrase “using only what the paper reports” is the load-bearing instruction. Without it, Anara fills gaps with background clinical knowledge, which is exactly what systematic review methodology prohibits. Restate it each session even if it feels redundant. Asking for the three quoted passages at the bottom of the response turns every extraction into a self-auditing unit: you are not reviewing the entry against the paper later, you are reviewing the quoted passages against the entry right now.

​
Check the output against your own understanding

Anara reads papers at the passage level and may miss data reported in tables, supplementary appendices, or figure legends. The outcome field is the most common failure point. If a primary outcome result looks wrong, check the paper’s tables and figures directly. For anything entering a published review, verify key quantitative fields against the source before they leave your log.

​
What to do with the output next

Paste each completed row into your extraction spreadsheet immediately. Reviewing the quoted passages is the quality step, and it is most useful right after extraction, not hours later. When your log is complete, use the follow-up comparison prompt to investigate outcomes where two papers diverge in ways your protocol did not anticipate.