Launching today
Equibles
MCP server for SEC filings, earnings calls & congress trades
5 followers
MCP server for SEC filings, earnings calls & congress trades
5 followers
An open-source MCP server that puts 1,000,000+ SEC filings and 100,000+ earnings-call transcripts inside Claude, ChatGPT and Gemini — every answer cited to the source filing, and engineered to refuse to guess. 90 tools + REST API over 5,000+ US stocks: filings, calls we capture and transcribe ourselves, company-specific KPIs read out of the filings (RPO, ARR, store counts), congress and insider trades, 13F, short data, macro. Free: 100 calls/day.
Hey everyone 👋
Ask any AI who NVIDIA's biggest institutional holder is and you'll get a confident answer stitched together from some 2023 blog post. The real data is public. It's just buried in 400-page filings, XBRL soup, and earnings calls that only exist as audio on a webcast that expires two weeks later.
I spent the last year building the pipeline that turns all of that into something an AI can actually query. Equibles gives Claude, ChatGPT and Gemini 90 tools via MCP (plus a REST API) over 1M+ SEC filings, 100k+ earnings calls, and 5,000+ US companies.
The part I'm proud of is I don't buy transcripts, the software gets the calls itself. The pipeline attends the webcasts across 13 different platforms, captures the audio, transcribes it, and matches every voice to the actual roster.
It also reads each company's own KPIs out of the filings such as RPO, ARR, ARPU, store counts, load factor. Those are the numbers that actually move a thesis, and no generic financials API carries them.
The ingestion engine is open source (AGPL on GitHub), so you can self-host it (the cloud version has extra tools). Free tier on the cloud is genuinely free: 100 calls a day, every tool, no card.
Built solo. I'd love to hear what data you'd want next, or anything that can help me improve the product. That's my roadmap.
- Daniel
Love how citations are baked in — feels safer than the usual "trust me bro" finance tools. One thing I'd kill for: a way to diff two earnings calls side by side so I can quickly see how management's tone and guidance shifted quarter over quarter.
@rsondur77420 Thank you! You can kind of do this already ask ALVIS AI "compare $NVDA's last two calls, how did tone and guidance change?" and it'll pull both transcripts and lay out the shift with citations. A. proper side-by-side view isn't built yet, though, but I really like that idea, putting it on the roadmap.
Really useful idea, would love to see a watchlist feature where I can save a set of tickers and get pinged when a new 8-K drops or when insider buys cluster for any of them.
@glenyulu20ou That's already supported! You can sign up and create e watchlist for free!
Pulled up a few 10-Ks on Equibles through Claude and the inline citations to exact filings are genuinely useful, not just vibes. The refusal-to-guess behavior caught me off guard in a good way when I asked something shaky.