Datavidence Financials for AI agents

Normalized US-GAAP statements for every SEC XBRL filer — point-in-time, as originally reported. Your AI agent calls them as tools over the Model Context Protocol, or you hit plain REST.

Try it — no signup GitHub Accuracy Main site

Why

Language models are good at reasoning about financial statements and bad at getting them. The data is all in SEC EDGAR, but EDGAR’s XBRL is a thicket of inconsistent tags, renamed concepts, and per-filer quirks — so agents either hallucinate the numbers or you spend weeks building a normalization layer. Datavidence Financials is that layer, delivered two ways: as a Model Context Protocol (MCP) server your agent calls as tools, and as a plain REST API. It’s published on the official MCP registry as ai.datavidence/datavidence-financials.

Try it in 30 seconds — no signup

FL_API_KEY=sandbox_demo_key uvx datavidence-financials

That runs the MCP server with a demo key that returns sample data — no account needed. Swap in your key for live figures.

Claude Desktop

Add this to your Claude Desktop config, then restart:

{
  "mcpServers": {
    "datavidence-financials": {
      "command": "uvx",
      "args": ["datavidence-financials"],
      "env": { "FL_API_KEY": "sandbox_demo_key" }
    }
  }
}

Then ask: “Get AAPL’s FY2023 income statement, balance sheet, and cash flow with ratios.”

The six tools

get_financialsNormalized income statement, balance sheet, and cash flow for a company & fiscal year — with optional as_of, provenance, and ratios.
get_financials_batchThe same for up to 25 companies in one call — built for peer comparison.
list_filingsDiscover a company’s available SEC filings and fiscal years.
get_revisionsHow a year’s figures changed across filings — as reported, then each restatement, with the filing date and EDGAR link behind every value.
search_companiesFind a ticker and CIK by company name, for when you have “Berkshire” and need BRK-B.
get_usageCheck your quota. Free to call — it doesn’t consume a request.

What makes it different

Point-in-time, as reportedPass an as_of date and you get the figures as they stood then — no look-ahead bias from later restatements. The thing backtests quietly get burned without.
Normalized to one schemaEDGAR’s XBRL tags drift between filers and change over the years. We map them to a single US-GAAP income statement, balance sheet, and cash flow — so you never parse a filing.
Every SEC XBRL filerCoverage is the whole SEC XBRL universe (~2009 to present), fetched and normalized on demand — any CIK with XBRL facts, not a curated subset.

Source-linked, too. Turn on provenance and every value carries its SEC accession, filed date, and a direct EDGAR URL — so the model cites the filing instead of inventing a number.

One known gap. When a company reorganizes under a new holding entity, SEC’s ticker map points only at the successor CIK — whose XBRL history starts at the reorganization. The earlier years stay under the predecessor CIK, so a ticker lookup for them comes back empty (the error names the range the successor actually covers). Reach the older filings by CIK instead: XOM resolves to the 2024 holding company, while FY2023 Exxon is cik=34088. Proper successor/predecessor resolution is on the roadmap.

Getting live data

The demo key returns sample data. For live figures, grab a key on RapidAPI:

Free $0 — 1,500 calls / mo
Pro $29 — 50,000 calls / mo
Business $99 — 300,000 calls / mo

Full feature parity across tiers — every tier returns the complete product (statements, ratios, point-in-time, provenance); they differ only by monthly call volume. Prefer plain HTTP? The same data is a REST call away — see the main site and API reference.

Where it’s at. Solo project, early, one instance. There’s no uptime SLA I’d promise yet. What I can say: the connector is open-source (MIT), every figure is traceable to its source filing, and I’m actively hunting for the cases where it’s wrong. If you build something with it — or it breaks — tell me.

Links

Registry ai.datavidence/datavidence-financials
Source github.com/datavidence/datavidence-financials-mcp
Package pypi.org/project/datavidence-financials
REST + keys RapidAPI

Data is derived from public SEC EDGAR filings on an “as is” basis, for informational purposes only — not investment, financial, legal, tax, or accounting advice. Verify figures against the original filings (use the provenance source_url). Sample/sandbox data is fictional, for integration testing only.