Introducing Headless Hebbia: your firm’s knowledge, wherever you use AI. Through our API and MCP server, we’re building bespoke applications with customers and bringing the same intelligence into the AI assistants their teams already use.

Introducing Headless Hebbia: your firm’s knowledge, wherever you use AI. Through our API and MCP server, we’re building bespoke applications with customers and bringing the same intelligence into the AI assistants their teams already use.

It's October, just in time for our THIRD major product launch these last few months! Introducing “Headless Hebbia” Our Matrix, Knowledge Graph, and Process Graph products now accessible via MCP or API... to power Claude, ChatGPT, Copilot, or internal builds with your institutional intelligence.

Hebbia will be at ITC this week in Las Vegas. Come find our team.

Opus 5.5 is a major step up over Opus 5. Our evals show it covers most of what our expert rubrics look for on end-to-end finance workflows. On retrieval evals, Opus 5.5 hit our best citation recall to date, with better token efficiency to keep cost in check. For teams running document-heavy research and citation-dependent workflows, this is the clearest jump we've measured between model versions.

In the span of a few weeks, AI has attacked one of math’s hardest problems, broken into systems built to contain it, and designed potential new cancer treatments. When will AI beat the markets? The next trillion dollar prize is staring the labs in the face...

Astra set a new high in our evals: it produced the best decks we've tested and followed the brief 17% more faithfully than the next-best model, while sourcing its claims to the right document 19% more often. For analysts working through dense financial materials, that means decks and answers you can hand to a client and defend line by line.

We had a busy summer of product ships. We launched Max, Matrix 2.0, and a wave of platform upgrades. Read the latest edition of The Disclosure: bit.ly/September_Disclosure

So PROUD of the @Hebbia team for today's Matrix 2.0 launch. Customer love grows every day... best feeling there is!

Great hosting so many finance professionals in our recent summer networking series. The Hebbia community keeps growing and we look forward to the next event.

made some merch. got a whole line loading...

Hebbia's George Sivulka on why token-maxing backfires: "If you want your employees to use AI, it makes sense to force them to hit some token leaderboard or overachieve on a token expectation. It's actually not a bad way to teach your organization how to use AI. But there's one of those interesting paradoxes where the minute you choose a metric to measure, it stops being a good indicator of success." "So now that we're measuring everyone's amount of tokens, people will use tokens for all the wrong reasons. They'll use AI for things they probably shouldn't, and in a really concrete way, harm the firm because they're spending a lot of the firm's resources for the wrong reason." "There's a canonical saying that 80% of the employees do 20% of the work, and vice versa. In the same way, you'll have lots of AI use that is completely non-productive, that doesn't actually improve the company or the organization itself." "And there's a very interesting implication where AI is now starting to evoke or create the use of more AI. So with looping agents or orchestrator systems, you can ask a simple question, and it can go and create more and more work, in the same way that human organizations create a lot of more work themselves." @gsivulka @hebbia

Claude Opus 5 is now available in Hebbia. Hebbia's Financial Services Benchmark shows meaningful gains in reasoning and citation accuracy over prior models. For investment banks, private equity firms, and credit funds, that means more reliable output on workflows like earnings analysis, comps, diligence, and more.