Document Lake stores your corporate documents — CIIAs, SAFEs, board consents, cap tables — organizes them automatically with AI, and answers questions about them with citations back to the source. Built for founders preparing for diligence, and for the moment someone asks “can you send me the signed agreement?”
$10 in AI credits included — no card required.
“Show me the CIIA John Doe signed.”
“What percent is the option pool?”
“In June 2025, what percent was the option pool?”
Drag in your corporate documents — CIIAs, SAFEs, certificates of incorporation, board consents, option grants, cap tables. PDF, Word, Excel, CSV, email, and more.
Each file is converted to clean markdown, classified, tagged by type/people/dates, and mined for dated facts like option pool percentages and signatories.
Search or chat in plain English. Answers cite the exact source document — and external AI agents can connect through the built-in MCP server.
Every document is classified and tagged on upload — by type, people, rounds, topics, and dates. Browse by facets, not file trees.
Documents become markdown + embeddings. Hybrid semantic and keyword search finds “the CIIA John Doe signed” on the first try.
Option pool, valuation caps, signatories — extracted as facts with effective dates, so “what was the pool in June 2025?” gets an exact, cited answer.
An AI assistant with read-only tools over your documents answers questions with [doc] citations — never from thin air.
Connect Claude, IDEs, or any MCP client over OAuth-protected Streamable HTTP and query the data room from your own agents.
Per-tenant isolation down to the storage layer, envelope encryption, append-only audit logs, EU data residency.
No subscriptions, no seat fees. On the hosted service you prepay AI credits and pay only for the AI you actually use.
$0 to start
Free forever
It's an AI-powered data room: a single organized home for a startup's corporate documents that founders, and the people they authorize, can search and question in plain English — for fundraising diligence, board prep, or just finding that one signed agreement.
No. Documents are processed only to serve your workspace (indexing, search, and the answers you request), under agreements with our AI providers that prohibit training on your content.
Always. Export your documents at any time, and deleting your workspace cryptographically destroys its encryption keys — the platform is also open source, so you can move to your own self-hosted instance.
Yes — the entire platform (the same code running this site) is AGPL-3.0 on GitHub. Self-host it on your own Cloudflare account with your own model keys, free.