From filing to answer.
Three steps turn official documents into knowledge your AI agents can trust, and that the community can open to everyone.
Official filing
Annual reports, notes, registry documents.
Knowledge graph
Companies, debt, subsidiaries and KPIs, linked.
Your agent
Asks a question, gets an answer with the exact page.
We read only official sources.
Annual reports, notes to the accounts and registry filings are collected and normalised. Nothing comes from press articles or scraped websites.
Every value keeps its coordinates: which document, which note, which page.
Your agent plugs straight in.
Our MCP server lets tools like Claude Desktop, Cursor or your own Python agents query the knowledge graph directly.
Instead of guessing, your agent walks the graph and returns the figure together with its source. Access is free, with fair usage limits.
The community chooses what comes next.
When a company is not yet covered, the community can fund a deep report on it. Once funded, the report is public for everyone and its data joins the graph.
Members get higher limits and monthly credits to fund reports, which keeps coverage growing and prices low.
Be first in line.
Join the waitlist to become a founding member, with bonus credits at launch.
Read the manifesto