canli-mcp 0.22.0 · free and open source

Give your AI real financial data.

canli-mcp is a free finance MCP server that connects Claude, ChatGPT, Cursor and other AI assistants to SEC filings, market and economic data and 443 finance tools. Every answer says where its numbers came from. No key, no account.

Add it to your assistantSee how it is tested

npx -y canli-mcp

Same question, asked twiceReceipt

AutoZone, Inc.: revenue for its fiscal year, from the 10-K it filed on

AI on its own
"not reported"
AI with canli-mcp
$20,338,555,000
Filed with the SEC
$20,338,555,000

Matches the filing

Open the 10-K on SEC.gov
gpt-6.1-terra, : one of 91 questions from that week's annual reports. The whole test.

What you can ask

Ask in plain words. It looks up the answer.

Your assistant picks the right tool and shows where each figure came from: the filing, the agency or the market data behind it.

More questions it answers

Works with

The assistant you already use

  • Claude
  • Claude Code
  • Claude Desktop
  • Cursor
  • VS Code
  • ChatGPT
  • Codex
  • Gemini CLI
  • Windsurf
  • Cline
  • Zed
  • Any other MCP client

Add it

One step, then ask

  1. In Claude or ChatGPT, as a connector

    Add a custom connector with this address. Nothing to install. The connector runs canli-mcp 0.22.0 and leaves out share prices and your own files; for those, and the newest release, run it on your computer.

    https://canlicapital.com/mcp/canli

  2. In Claude Code

    claude mcp add canli -- npx -y canli-mcp

  3. In every AI app on your computer at once

    It finds Claude Desktop, Cursor, VS Code and the others, adds one entry to each and keeps a copy of every file it changes.

    npx -y canli-mcp setup

Step-by-step for each app

What is inside

Seven toolkits in one

Each is also its own server. canli-mcp loads them together, so your assistant reaches all 443 tools through one connection.

  • 236 tools

    Quant finance

    performance and risk, options, fixed income, portfolios, econometrics, indicators and a library of 399 strategy sleeves

    About quant finance

  • 17 tools

    Backtest validation

    deflated Sharpe ratio, overfitting probability, data-snooping tests, leakage checks and placebo tests

    About backtest validation

  • 139 tools

    Markets

    SEC filings and statements as filed, foreign filers too, insider trades, 13F and 5% holders, fund portfolios, options chains and volatility surfaces, the Fed, Treasury yields and auctions, FRED, World Bank and OECD data, central banks abroad, prediction markets, prices and a whole-market overview, analysts' targets and technical indicators, a figure traced down its filing, what a token's contract can do, every annual report searched at once, management's record against its own guidance, what followed every past market fall, ship traffic from satellite tracking, every country's economy from its own statistics office and central bank, the Fed's next moves as futures price them, nowcasts of growth and inflation, money markets and financial conditions, market breadth and the earnings season, and energy, crops, jobs and housing from their publishers

    About markets

  • 8 tools

    SEC fundamentals

    company financials as first reported, with later restatements flagged

    About sec fundamentals

  • 6 tools

    Research record

    Canli Capital's open research record

    About research record

  • 5 tools

    Paper trading

    Alpaca paper orders behind pre-trade checks, each check signed

    About paper trading

  • 32 tools

    Factor backtests

    factor backtests that refuse lookahead and fill only what the market can take, crisis replays, purged cross-validation, a breadth lab that counts its independent bets, parameter sweeps on every core, data annotation, a signed ledger of filings, a paper league, a paper exchange where AI agents trade in order books, a count of how many independent views a group of models holds, and a search for what diversifies a book

    About factor backtests

Every tool, what it does and what it takes

It can draw, too

Charts in the chat, for a short summary's cost

In apps that show MCP views, the result opens as a chart or table you can sort and zoom, while the assistant reads only a short summary.

How it is tested

Checked against the filings, every week

  • Every week

    Questions about last week's filings

    gpt-6.1-terra alone answered 0 of 91; with canli-mcp, 89 of 91.

    The weekly test

  • Against other finance servers

    79% right on the 14 questions every server finished; the best of the others, 54%.

    The comparison

  • Every release

    How it works and is tested

    Each tool against reference cases, every search for a tool, every number's source.

    How it works

Questions

Before you add it

Is it free?

Yes. canli-mcp is open source under the MIT licence. It needs no API key and no account, on your computer or as a connector.

Does it give investment advice?

No. It shows what companies filed and what the data says, with the source of each figure. What you do with it is your decision.

Where does the data come from?

From the people who publish it: the SEC's filings and their XBRL data, the Federal Reserve, the Treasury, the Bureau of Labor Statistics, FRED and others. Prices come from Yahoo Finance's public chart data unless you add your own Alpaca or Tiingo key.

What do I need to run it?

As a connector, nothing: paste the URL into Claude or ChatGPT. On your own computer, Node.js 20.10 or later. One command adds it to every AI app it finds.

Who makes it?

Arhan Canli, 15, in Dubai, in the open on GitHub. Corrections are published on the corrections page.

The family

Every Canli server

For developers: a local paper-trading workflow, not yet released. Not investment advice. The hosted servers can run an earlier release than npm; each card shows its own.