Every quarter, the ritual repeats itself: hundreds of hedge fund managers sit down to explain—in narrative form—exactly why they bought what they bought, what worries them, and where they see value hiding. These letters, thick with conviction and occasionally hubris, get published as PDFs. Then they disappear into the digital ether, useful primarily to investors patient enough to hunt through prose for a single company mention.
It's a strange inefficiency, really. The investment world runs on data, yet some of its richest material lives locked in documents that resist the kind of mass querying that defines modern finance.
Canadian fintech Fiscal.ai thinks it's found a way through. The company has announced a new API product that ingests more than 7,000 investor letters—spanning 700-plus global funds—and converts them into something resembling structured data. Not the usual financials-and-estimates fare. This is messier: what investors actually write about the companies they own, complete with thesis details, conviction indicators, and catalysts they're banking on.
Breaking Open the Black Box
Here's how it works. Fiscal.ai pulls quarterly and annual letters from hedge funds, mutual funds, and partnerships dating back to 2020. Its system scans each document for company mentions, then extracts and structures what it finds: bullish, bearish, or neutral stance. Conviction level. Position sizing clues. Summaries of the underlying thesis, plus identified risks, expected catalysts, valuation arguments. A representative quote from the original text, for flavor.
Each structured "thesis object" links back to the source PDF—provenance matters when you're turning narrative into queryable fields. The company says it curates the corpus actively, deduplicating letters that show up across multiple channels and vetting for completeness. Users can query by company, investor, time period, or topic through a set of API endpoints. One call returns all published theses about a single stock across different managers. Another surfaces investor profiles with coverage stats and inferred strategy styles.
The product is live now across Fiscal.ai's API tiers. There's a free trial (capped at 45 companies and 250 calls daily), and pricing runs up to $199 monthly for Enterprise access.
A Crowded Lane, Suddenly
Fiscal.ai isn't pioneering virgin territory here. Hedge Fund Alpha has launched its own structured letter dataset, mining a similar archive. Free alternatives like BuySide Digest and HF Best Ideas offer searchable databases of fund letters, though they lean toward discovery and topical indexing rather than programmatic access.
The broader picture? The financial data industry is in a sprint to structure qualitative narratives. S&P Capital IQ Pro has been layering GenAI-powered document intelligence into its platform to analyze filings and transcripts. Quill AI, backed by Y Combinator, extracts equity research from unstructured regulatory filings. Chronograph rolled out semantic search for private-market portfolio documents back in March 2026.
The pattern is unmistakable: narrative is the new frontier, and everyone's racing to parse it.
The Pitch, and the Gaps
For asset managers, the use case feels intuitive enough. Scan fund theses by ticker to see which managers like a name and why. Track shifts in narrative framing from one quarter to the next. Map investor style alignment when hunting for ideas. It's faster than manual letter-reading marathons, and it adds context that pure quant signals miss.
Fiscal.ai—which rebranded from FinChat in 2025—targets both individual and institutional buyers, including platform operators building atop the API. The company raised a $10 million Series A led by Portage in 2025 and has partnerships with firms like YCharts for distribution.
But there are blanks. The 700-investor, 7,000-letter corpus? That's a company claim. No independent verification yet. No disclosed breakdown of which funds made the cut and which didn't. The API documentation is thorough on structure and fields, less so on accuracy benchmarks or how the system navigates ambiguous language—hedge fund letters aren't always models of clarity.
Still. Investor letters contain something filings and transcripts don't: thesis-level color. Why a manager bought in, what they're nervous about, which catalyst they're waiting for. Fiscal.ai is betting that making that information queryable at scale justifies the engineering effort.
Given how many competitors are suddenly moving into adjacent space, the bet doesn't look unreasonable.
