Canli Research — Claude Code MCP server: what it does, how to install
Provides read-only access to Canli Capital's quant research library, failed trial counts, and simulated trading records. Helps agents inspect quantitative hypotheses before building strategies.
claude mcp add --transport http canli-research-mcp https://canlicapital.com/mcp/researchWhat Canli Research does
The MCP server exposes six read-only tools to your coding agent. The search_research tool queries published papers by keyword, while get_paper retrieves specific article text and sections. Discarded quant ideas can be audited against the research budget using trial_ledger.
The server fetches static files directly from the canlicapital.com CDN. It caches responses in memory for ten minutes. A repeated inquiry requires no network roundtrip. You do not need API tokens, cookies, or database configurations.
Every published metric reflects paper trading execution rather than funded live money, meaning results do not serve as investment advice. The process writes nothing to local storage.
What it helps with
- review quantitative ideas against previously discarded hypotheses
- search public investment research by keyword and topic
- fetch paper text or specific sections into model context
- inspect live paper-trading records and sleeve equity
- verify the head of the tamper-evident audit chain
Example prompts for the agent
- Run search_research to see if Canli Capital has published studies on trend following.
- Check trial_ledger to inspect how many candidate hypotheses were killed during testing.
- Use prompt preflight_idea to evaluate whether this statistical arbitrage concept was tried before.
How to install Canli Research
1Platform CLI: remote server
claude mcp add --transport http canli-research-mcp https://canlicapital.com/mcp/research2Platform CLI
claude mcp add mcp-research -- npx -y canli-research-mcp1Platform CLI: remote server
codex mcp add mcp-research --url https://canlicapital.com/mcp/research2Platform CLI
codex mcp add mcp-research -- npx -y canli-research-mcpConfig file
{
"mcpServers": {
"mcp-research": {
"url": "https://canlicapital.com/mcp/research"
}
}
}{
"mcpServers": {
"mcp-research": {
"command": "npx",
"args": [
"-y",
"canli-research-mcp"
]
}
}
}Config file
{
"servers": {
"mcp-research": {
"type": "http",
"url": "https://canlicapital.com/mcp/research"
}
}
}{
"servers": {
"mcp-research": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"canli-research-mcp"
]
}
}
}1Platform CLI: remote server
gemini mcp add -t http mcp-research https://canlicapital.com/mcp/research2Platform CLI
gemini mcp add mcp-research npx -y canli-research-mcpStep by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about Canli Research
How to install Canli Research in Claude Code?
Run claude mcp add --transport http canli-research-mcp https://canlicapital.com/mcp/research in your terminal. This connects your agent directly to the hosted endpoint without local build steps.
What does Canli Research do?
It gives coding assistants read-only access to Canli Capital's research catalog. Agents can read published papers, track rejected hypotheses, and review simulated portfolio sleeves.
Does Canli Research require an API key?
No API key or credentials are required. The server queries only public CDN assets and stores no data on your local filesystem.
