Quantplay MCP server
Let Claude, ChatGPT or any MCP client check positions, read option chains, place orders and set limits on 10+ Indian brokers. One URL, OAuth sign-in, no API key.
Last updated 21 October 2026
On this page
What it is
The Model Context Protocol (MCP) is an open standard for connecting AI assistants to outside services. The Quantplay MCP server lets Claude, ChatGPT, Cursor and other MCP clients work with the broker accounts you have connected to Quantplay. For other ways to connect, including OpenClaw and Hermes, see AI agents.
What you can do
- Read positions, holdings, orders, margins, live P&L and history
- Look up option chains, expiries, instruments and market data
- Place, cancel and square off orders, if you allow trading
- Manage copy-trading connections
- Set stoploss, target and trailing stoploss, if you allow risk changes
Authentication
No API key needed. Add the URL, pick OAuth if asked, then sign in and choose what the agent can do. Viewing is always on; trading and risk changes are off until you allow them.
Full access
https://mcp.quantplay.tech/mcpRead-only
Can't place, change or cancel orders, even if trading was approved.
https://mcp.quantplay.tech/mcp/readonlySetup guides
Claude
Requires a Claude plan with custom connectors.
- 1Copy the Full access URL above.
- 2Open Settings → Connectors → Add custom connector, paste the URL and click Add.
- 3Click Connect. Sign in to Quantplay, choose permissions and approve.
- 4Start a new chat and ask for your Quantplay accounts to check it works.
ChatGPT
Requires a ChatGPT plan with Developer mode.
- 1Copy the Full access URL above.
- 2In Settings, open Apps → Advanced settings and turn on Developer mode.
- 3Create an app, paste the URL and choose OAuth.
- 4Sign in to Quantplay, choose permissions and approve. Then ask ChatGPT for your Quantplay accounts.
Claude Code
- 1Run:
claude mcp add --transport http quantplay https://mcp.quantplay.tech/mcp - 2Optional: add
--scope userfor every project, or--scope projectto share it through.mcp.json. - 3Run
/mcp, pick quantplay and choose Authenticate.
Codex
- 1Run:
codex mcp add quantplay --url https://mcp.quantplay.tech/mcp - 2Run
codex mcp login quantplayand approve the sign-in in your browser.
Cursor
- 1Add this to
~/.cursor/mcp.jsonfor every project, or.cursor/mcp.jsonfor one:{ "mcpServers": { "quantplay": { "url": "https://mcp.quantplay.tech/mcp" } } } - 2Approve the sign-in Cursor opens.
- 3Check the tools under Settings → MCP.
Any other MCP client
- 1Add the Full access URL as a remote server using Streamable HTTP.
- 2Pick OAuth and approve the sign-in.
- 3No browser? Send an API key from Settings → Developer as an
X-Api-Keyheader. A key can do everything your account can, so prefer OAuth.
Example prompts
Your risk limits still apply
Stoploss, target and kill switch are set per broker account and run on Quantplay's servers. When a limit is hit, Quantplay exits the agent's positions just as it exits yours. No agent can change your kill switch.
Technical details
Troubleshooting
Asked to sign in again and again
Finish the approval step in the Quantplay tab that opens. If it keeps happening, remove the connector, add it again and approve once more.
Using an old URL with a key in it
URLs with ?QUANTPLAY_API_KEY= no longer work. Use the plain URL and sign in with OAuth, which keeps no credential in your client config.
Tools not showing up
Restart your client after changing its config, check your JSON for typos, and confirm the URL. Server status: https://mcp.quantplay.tech/health.
Disconnecting an agent
Settings → AI Integration → Disconnect in Quantplay. Access ends within a minute.
Still stuck?
Email hello@quantplay.tech or WhatsApp us from the support page. What we store when you connect an agent is in our Privacy Policy.