An answer
you cannot trace
is a rumour.
Docs.
Cited.
Yours.
Open-source, self-hosted AI document intelligence. Ask across your own archive and get answers grounded in cited sources.
Documents are everywhere.
Knowledge is not.
Scattered
Drive, Dropbox, Slack, email, a share nobody has opened in a year. People remember the content, not the location.
Hard to retrieve
Keyword search matches filenames and strings. It does not cross the silos, and it does not know what you meant.
Outside the model
A general-purpose AI knows the public internet. It has never read your contracts, your policies, or last quarter’s memo.
Can’t leave the building
For legal, healthcare, finance and government archives, moving the documents into someone else’s platform ends the conversation.
Your archive
becomes queryable.
Point AskDocs at the documents you already have and ask in plain language. Every answer is assembled from passages it retrieved — and every passage is a citation you can open.
Documents
Point it at the archive you already have. Nothing migrates, nothing re-formats.
Question
Ask in plain language. No query syntax, no tags to remember.
Cited answer
Open every passage the answer was assembled from.
Your documents.
Your infrastructure.
Your answers.
Open source under MIT. Clone the repository, set two values, and bring the stack up.