Document Intelligence
Evidence-first document intelligence
Ask your documents. Verify every answer.
Turn a document into focused answers, then inspect the retrieved passages and citations behind each conclusion.
What is DocSage?
Upload a document — a résumé, report, contract, or research paper — and DocSage uses retrieval-augmented generation (RAG) to extract a structured assessment: overall fit, strengths, gaps, risk factors, and focus areas.
01 — Ingest
Drop in a PDF, Markdown, HTML, or plain-text file. The document is chunked and embedded into a pgvector index.
02 — Retrieve
Your query is embedded and the top-k most semantically similar chunks are retrieved using cosine similarity.
03 — Analyse
An OpenAI LLM synthesises the retrieved chunks into a structured report. Every claim is grounded in source citations.
Ingest
Explore before you upload
Inspect a precomputed analysis of a fictional document. No file, database record, embedding, or model call is created.
The original file is processed in memory and is not retained.
Extracted text, chunks, and embeddings are stored for retrieval.
Do not upload confidential material to this public demo.
ACCEPTS PDF · MD · HTML · TXT — max 10 MB
Analysis
— upload a document to begin —