A view is a table built from a folder. Each row is one of your files, each column is a question, and Anara fills in every cell by reading that file and answering that question, with a citation behind every answer.
A folder of papers usually holds the same handful of facts in every file: the sample size, the method, the outcome, the limitation. The usual way to get them out is to open each paper, ask it the same question, and copy the answers into a spreadsheet by hand. A view does that pass for you. Ask once and every file answers.
What a view is
Every folder starts with the Default view, the plain list of files you already know. A view you create keeps those rows, one per file, and adds columns. Each column is one question, and you can set whatever questions you want.
Describe what you want to know and Anara writes the columns for you, then works through the table cell by cell. For the first cell it reads the first file and answers the first question, and it keeps going until the table is full.
Cells are not summaries. Each one is a specific value taken from the file, with the passages it came from attached, so you can check any answer in a click.
A folder can hold as many views as you like, and switching between them never changes the files themselves.
When to use a view
Reach for a view when you want the same question answered for every file:
Literature review extraction: sample size, population, design, effect size, and limitations across every paper at once
Screening: pull the two or three fields you need to decide what is in scope
Comparing methods or results between studies side by side
Building a summary table for a paper or report, sourced from the originals
If your question is open ended, or you want one synthesized answer rather than a value per file, use folder chat instead.
How to create a view
Open a folder in your library.
Click View in the toolbar above the table, then click Create view.
Describe what you want to extract. Suggested prompts appear above the input, written from what is actually in the folder, so you can click one and edit it rather than starting from a blank box.
Click the arrow to start.
Anara writes the columns from your description, names the view, and starts filling cells. You will see Building your extraction table while it sets up, then Extracting in each cell as it works. Results land column by column.
To build a view from part of a folder rather than all of it, select the files you want with the checkboxes, then click Extract to table.
Tip: Describe the fields, not the task. "Extract sample size, study design, primary outcome, and limitations" gives you far cleaner columns than "review these papers".
Creating and editing views from chat
You can do all of this by asking instead. In a folder chat, ask for something like "create a table extracting the country of the first author and the funding source from this folder". Anara builds the view and starts the extraction.
If you already have a view open, the same kind of request applies to that view: ask it to add a column, change what a column looks for, rerun it, rename it, or delete it.
How columns work
Each column is one question, asked of every file.
To add one, click Add extraction column at the right end of the header row, then write your question in the instruction box, for example "What was the sample size?". If you leave the name as New column, Anara names the column for you from the question you wrote.
Set Type to control the shape of the answer: Text, Number, Date, Yes / no, or Single select. Leave AI citations on so every answer keeps its sources.
A new column starts extracting the moment you click Save column. To change one that already exists, open its header menu for Edit extraction column, Retry column, and Delete column. Changing what a column looks for re-runs that column.
Checking every answer
Hover a filled cell and a small pill appears showing how many passages the answer came from. Click it to open Source evidence: the file, the page number, and the exact text Anara used. Click any of those rows to open the file at that page with the passage highlighted.
Click the cell itself to read the whole value when it runs longer than the column is wide.
This is the point of a view. You are never asked to take the table on trust, because every cell traces back to the sentence it came from. Citations work the same way here as they do in chat.
Keeping a view up to date
A view built from a whole folder follows that folder. Add files and their rows appear and fill in on their own. Remove files and their rows go with them.
To re-run everything, including cells that are already filled, click Rerun in the toolbar. Beside it you will see a running count of any cells that failed or came back empty.
Which files a view can read
Views read documents, websites, spreadsheets, notes, videos, YouTube links, audio, and recordings. Images are not supported and are skipped.
If a file is still importing when a view runs, its cells wait rather than fail, and they fill in once the import finishes.
Note: One run fills up to 2,000 cells, counting files multiplied by columns. If a view comes out larger than that, run it with fewer columns or over a smaller set of files.
If something isn't working
Cells show Failed: usually the file itself cannot be read. Open it and check that it loaded properly, since a file that shows File unavailable in the viewer has no text to extract from. Click Retry on the cell once the file is fixed.
Cells show Not reported: Anara read the file and the answer genuinely was not in it. That is a real result rather than an error. If you expected something there, reword the column to match the vocabulary the paper uses.
Cells show Waiting: the file is still importing, and they will fill in on their own.
A column keeps answering the wrong question: edit that one column's instruction to be more specific instead of rebuilding the whole view.
Still stuck? Here's how to get support.