When you upload or import a document, Anara does more than store the file. It extracts the full text, indexes it for search, and makes every page available for AI analysis. Understanding how this works — and what can go wrong — helps you get the most out of your library.
What happens when you upload a document
Think of processing like a librarian reading a new book before shelving it. After a file is uploaded, Anara processes it in four steps:
Text extraction — Anara reads the document and extracts the text content from each page.
Metadata extraction — Title, authors, DOI, publication, and other fields are identified and stored.
AI summary — An AI-generated summary is created from the extracted text.
Indexing — The text is indexed so the AI can search it when you ask questions.
This happens once after upload. Once complete, the AI has access to the full content of every page — not just the abstract, not a preview, but the actual text.
Full text statuses
In your library table, each document shows a Full text status. This tells you how much of the document is available to the AI:
Full text: Available — Anara has extracted the complete text. The AI can read and cite any part of the document. This is the ideal state for research.
Abstract only — Anara found the metadata but could not access the full PDF. This typically happens with paywalled papers imported by DOI. The AI can answer questions based on the abstract but cannot cite specific passages from the body.
Content unavailable — The full text and abstract could not be retrieved. This usually happens when a paper is behind a paywall and was imported by DOI or URL. The AI has limited ability to help with these documents.
Other statuses you may see: Website (imported web pages), Video (video imports), and Note (notes you wrote in Anara).
Tip: If a document shows "Content unavailable", try downloading the full PDF through your institution or library and uploading it directly. This is the most reliable way to ensure Anara has full access to the text.
Getting better results from your documents
Import by DOI or URL when possible. Anara can often retrieve the full text from the source when given a direct identifier.
If a paper shows "Abstract only", find the full PDF from the publisher or your institution and upload it directly. This gives the AI full access to the paper.
For papers behind paywalls, look for an open-access version through your library's digital access or a repository like PubMed Central.
What the AI does with your document
When you ask a question in the viewer sidebar, the AI searches the indexed text for relevant passages, synthesizes an answer, and cites the specific passages it used. It is working from the actual text of your document — not a surface-level description.
The more complete the text extraction, the more accurate and specific the AI's answers will be. For documents with "Full text: Available", the AI can reason across all pages — conclusions, supplementary data, methodology, any section.
Tip: If a document shows "Abstract only" or "Content unavailable", the AI will say so rather than guess. This is intentional — Anara does not generate answers beyond what it can verify in your documents.