The full upstream README, mirrored here for reference. Install config, tool schemas, adoption signals, and an original overview live on the BlindPay listing page.

The official Model Context Protocol server for BlindPay - Stablecoin API for global payments.
This MCP server provides AI assistants (Cursor, Claude Code, Codex, etc.) with access to BlindPay's stablecoin payment infrastructure. It exposes tools that allow AI assistants to:
| Remote server (recommended) | Local npx server | |
|---|---|---|
| URL / command | https://mcp.blindpay.com/mcp | npx -y @blindpay/mcp |
| Auth | OAuth 2.1 sign-in with your BlindPay account (PKCE) | BLINDPAY_API_KEY env var |
| Transport | Streamable HTTP | stdio |
| Works in | Claude, ChatGPT, Cursor, VS Code, Grok, any remote-capable client | Any stdio MCP client |
Both expose the same tools. The remote server needs no API key: the client opens a browser, you sign in and pick an instance, and the token is scoped to your role on it.
Set BLINDPAY_MCP_PROFILE on the local server to choose which tools are exposed:
| Profile | Tools | Use when |
|---|---|---|
full (default) | Every public API operation, including payouts and transfers | Building or operating with an AI coding assistant |
readonly | GET operations only: status, balances, quotes, history | Consumer assistants, or any place a tool that moves money is unwanted |
Every tool carries MCP annotations (readOnlyHint, destructiveHint, idempotentHint) derived from its HTTP verb, so clients can gate confirmations themselves.
Get your API key and Instance ID:
in_xxxxxxxxxxxx) from the dashboard urlDependencies you need to have installed:
You can check if you have Node.js installed by running node -v. Same for Npm, you can check by running npm -v.
One-click install (remote server):
One-click install (local server with API key):
After installation, add your API key and instance ID to ~/.cursor/mcp.json:
As a plugin (remote server, OAuth, no API key):
Or as a plain MCP server:
Local server with an API key:
Add to your Claude Desktop configuration file:
~/Library/Application Support/Claude/claude_desktop_config.json%APPDATA%\Claude\claude_desktop_config.jsonRemote server (OAuth, no API key):
Local server with an API key:
Or add to your ~/.codex/config.toml (or project-scoped .codex/config.toml):
Click the Install in VS Code badge at the top, or add to .vscode/mcp.json:
You can also search @mcp blindpay in the Extensions view once the gallery entry is live.
Add to ~/.codeium/windsurf/mcp_config.json:
Open the Model Context Protocol Registry extension in Raycast, search for BlindPay and click Install Server. Or install manually with Manage Servers → Add Server using the local config from the Cursor section.
In Settings → Connectors → Add custom connector, paste https://mcp.blindpay.com/mcp and complete the sign-in. Only read-only tools are exposed to consumer directories.
Enable Developer Mode in Settings → Connectors → Advanced, then Create a connector with https://mcp.blindpay.com/mcp and OAuth authentication.
Grok Build plugin (recommended): install blindpay from the xAI plugin marketplace. The plugin ships both remote HTTP MCP endpoints:
| Server | URL | Scope |
|---|---|---|
blindpay | https://mcp.blindpay.com/mcp | Full (read + write) |
blindpay-readonly | https://mcp.blindpay.com/mcp/readonly | Read-only |
Both use OAuth 2.1 (PKCE) with your BlindPay account — no API key to paste.
Custom connector (paid Grok tier): go to https://grok.com/connectors → New Connector → Custom, enter https://mcp.blindpay.com/mcp (or the readonly URL), leave client credentials empty, and complete the BlindPay sign-in when prompted.
Local stdio server in a container (also listed in the Docker MCP Registry):
Build it yourself with docker build -t blindpay-mcp . from this repo.
Remote: point the client at https://mcp.blindpay.com/mcp (Streamable HTTP, OAuth 2.1 discovery via /.well-known/oauth-protected-resource).
Local: run npx -y @blindpay/mcp over stdio with BLINDPAY_API_KEY set.
Once configured, you can ask your AI assistant to interact with BlindPay:
[!TIP] You can specify
BLINDPAY_INSTANCE_IDin your initial prompt or add it as an environment variable. However, some AI tools may not detect it automatically, so you may need to include it explicitly in your prompt.
This project is licensed under the MIT License - see the LICENSE file for details.
Made with ❤️ by the BlindPay team