Signals.ai

Companies say more than they file. Signals.AI covers both for every US-listed company: the SEC filings, and the earnings calls, slide decks, press releases and investor day material companies publish themselves. Deep research on a company. Business model, financials, ownership, risks and competition, assembled end to end and exported as a cited PDF. Operational KPIs. The numbers management reports but nobody files in a standard format. Backlog. Remaining performance obligations. Annual recurring revenue. Net revenue retention. Same-store sales. Segment revenue. Pulled from company documents and earnings call transcripts. Ownership and insider activity. Institutional positions from 13F, insider trades with cluster detection, and congressional trading disclosures. Earnings analysis. What changed this quarter. Guidance revisions, prepared remarks versus Q&A sentiment, and whether management is still saying what it said last quarter. Live transcripts stream speaker-separated turns while the call is happening, with a finalized, cited transcript within minutes of the call ending. Screening and discovery. A screener that takes plain English, so "profitable software companies with recent insider buying" returns a ranked list with reasons, plus thematic exploration and scenario modeling. All of it runs on one data layer. From regulators: SEC filings, fundamentals, insider transactions, 13F holdings, congressional trades, short interest, and ETF and mutual fund holdings. From companies: earnings call transcripts, press releases, earnings slide decks, investor day presentations and fact sheets, pulled straight from investor relations sites. Slide decks are never filed with the SEC, which is why they are missing from most research tools. Every figure links back to the document it came from, so anything the system tells you can be verified against the original. Data refreshes as companies file and report, so screens and alerts reflect what landed today rather than last quarter. The same first-party corpus is available programmatically. A versioned REST API covers filings, fundamentals, KPIs, transcripts, IR documents, ownership, insider, congress and short interest. A read-only MCP endpoint lets AI agents such as Claude, Cursor and ChatGPT query KPI history, transcripts and IR documents mid-conversation with no connector to write. Keys are self-serve, with documentation and examples at docs.signals.ai. Much of the product is free with no account: