CNEquity
Read-only access to a local China A-share data lake for stock research
- Pricing
- Open source
- Platforms
- CLI, self-hosted, macOS
- Runs
- Locally
- Version
- 0.21.0
- Listing updated
- 3 days ago
About CNEquity
CNEquity is an open-source MCP server for finance and payments. It runs locally on your machine and is started by your AI client.
CNEquity keeps 86 datasets of China A-share data as local Parquet files, and its cne mcp command lets an AI client query them without any way to trigger collection, retries or cleanup. Data covers prices, financial statements, capital flows, company events and futures, with delisted stocks retained and financial reports kept point-in-time by announcement date.
The server offers six tools, split by the shape of the question rather than by dataset. Correctness rules travel inside responses: price queries without an adjustment setting return a warning, and fundamentals queries refuse to run without an as-of date, since defaulting to today would answer a historical question with later information. Heavy reads above four million rows without a narrow symbol list are rejected, and the response names the latest date window that fits.
It works over stdio with Claude Code, Claude Desktop, Codex, Gemini CLI, Cursor, VS Code and others. ChatGPT needs an HTTP mode with a token and an HTTPS tunnel. A live mode fetches data on demand without writing it to disk. The project is 0.x, has no backtesting engine or order placement, and the code is Apache-2.0 while the data stays under upstream licences.
Install
Run in a terminal
claude mcp add cnequity -- uvx cnequityAdd to claude_desktop_config.json
{
"mcpServers": {
"cnequity": {
"command": "uvx",
"args": [
"cnequity"
]
}
}
}Add to ~/.cursor/mcp.json
{
"mcpServers": {
"cnequity": {
"command": "uvx",
"args": [
"cnequity"
]
}
}
}Add to .vscode/mcp.json
{
"servers": {
"cnequity": {
"type": "stdio",
"command": "uvx",
"args": [
"cnequity"
]
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"cnequity": {
"command": "uvx",
"args": [
"cnequity"
]
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"cnequity": {
"command": "uvx",
"args": [
"cnequity"
]
}
}
}Tools this server provides (6)
Taken from the project documentation; the running server may differ.
- describe_lake
- Reports what the lake contains, its coverage and the conventions that make answers correct.
- resolve_symbol
- Maps a company name to its ticker, including delisted stocks with a delist date.
- query_bars
- Reads daily, index and minute bars with adjust and universe options.
- query_fundamentals
- Reads financial statement items; requires an as-of date for point-in-time results.
- query_dataset
- Reads any other dataset filtered by date and symbol.
- run_sql
- Runs a single read-only DuckDB SELECT for aggregation and ranking.
What it does
- Six question-shaped tools
- Describe, resolve symbols, bars, fundamentals, other datasets and SQL, instead of one tool per dataset.
- Point-in-time fundamentals
- Financial statement queries require an as-of date to avoid look-ahead bias.
- Contract in responses
- Adjustment, universe and coverage rules are returned with results, not just documented.
- Row limits with guidance
- Oversized reads are refused and the response suggests a window that fits.
- Delisted stocks kept
- Historical universes include delisted names, flagged with a delist date.
- Live mode
- Fetch data on demand without building a local lake.
Good for
- Ask for the highest closing price of a stock over recent trading days
- Look up which financial data was known on a rebalance date
- Rank sectors by capital flow using a DuckDB query
- Resolve a Chinese company name to its ticker
- Prototype quant research before building a full local lake
Pricing
Open source
Apache-2.0 code; data is subject to upstream licences
Prices change. Check rootsunc.github.io before you buy.
CNEquity FAQ
Can the AI trigger data collection through MCP?
No. The MCP interface is read-only and has no entry point for collection, retries or cleanup.
How do I try it without building the full lake?
Run cne init --profile demo to fetch a small slice of real data, or use the --live option.
Does it work with ChatGPT?
Yes, in HTTP mode with a token and an HTTPS tunnel, and it needs developer mode on a paid plan.
Is there a backtesting engine?
No. The project does not include backtesting or order placement.
Is CNEquity free?
Yes. CNEquity is open source, so you can use and self-host it for free.
How do I add CNEquity to Claude Code?
Run this command in your terminal: claude mcp add cnequity -- uvx cnequity
What are the best alternatives to CNEquity?
The closest alternatives to CNEquity are Agent-Native Analytics, Google Ads + Meta Ads + SEO/GEO MCP, Solana MCP by Vybe, Tapetide - Indian Stock Market MCP and World Monitor.
Alternatives to CNEquity

Agent-Native Analytics
MCP serverOpen-source analytics app where an AI agent builds charts and dashboards from your data
- Remote
- Open source
- 7.7K GitHub stars
- Analytics and data

Google Ads + Meta Ads + SEO/GEO MCP
MCP serverHosted MCP connector for managing Google Ads, Meta Ads, GA4 and Search Console from chat
- 8 tools
- Remote
- Freemium
- 8.1K GitHub stars
- Marketing and SEO

Solana MCP by Vybe
MCP serverRemote MCP server exposing the Vybe Solana API to AI clients and code editors
- 44 tools
- Remote
- 2.8K GitHub stars
- Finance and payments

Tapetide - Indian Stock Market MCP
MCP serverIndian NSE and BSE stock market data for AI assistants through 48 MCP tools
- 17 tools
- Remote
- Freemium
- 150 GitHub stars
- Finance and payments

World Monitor
MCP serverLive geopolitical intelligence dashboard with an API and MCP server for agents
- Remote
- Freemium
- Analytics and data
Budget Tracker
MCP serverLocal MCP expense ledger that logs receipts and mileage, splits VAT and prepares rebilling
- 14 tools
- Remote
- Freemium
- 2 GitHub stars
- Finance and payments