windsurf-search-mcp
MCP server and CLI for Windsurf/Devin server-side web search, exposing a web_search tool that returns JSON hits from the GetWebSearchResults API.
README
windsurf-search-mcp
MCP server and CLI for Windsurf/Devin server-side web search
(GetWebSearchResults).
Zero runtime dependencies. Node.js >= 20.
This talks to Windsurf/Devin cloud endpoints with a personal session token. Use at your own risk; tokens expire and may violate the provider's terms if used outside the official client.
Install
npm i -g windsurf-search-mcp
# or without install:
npx -y windsurf-search-mcp --help
Auth (no secrets in config files)
Resolve order:
--api-key <token>WINDSURF_API_KEY(or legacyWINDSURFAPI_CODEIUM_API_KEY)- first existing key file:
~/.config/windsurf-search/api-key~/.windsurf-search/api-key~/.piwin/windsurf-api-key(compat)
Expected token shape: devin-session-token$...
# interactive (masked)
windsurf-search config set
# or non-interactive
windsurf-search config set 'devin-session-token$...'
windsurf-search config show
windsurf-search config test
Email/password login is also available (windsurf-search --login), but many
accounts are OAuth-only and will reject password login.
CLI
windsurf-search "tauri window drag region" --limit 5
# stdout JSON:
# { "hits": [ { "title", "url", "snippet", "source": "windsurf" } ] }
Useful for agent hosts that spawn a custom CLI search source and parse JSON hits.
MCP server
Cursor / Claude Desktop / generic MCP host
{
"mcpServers": {
"windsurf-search": {
"command": "npx",
"args": ["-y", "windsurf-search-mcp"],
"env": {
"WINDSURF_API_KEY": "devin-session-token$..."
}
}
}
}
Prefer putting the token in a key file and omitting env entirely:
{
"mcpServers": {
"windsurf-search": {
"command": "npx",
"args": ["-y", "windsurf-search-mcp"]
}
}
}
Then:
windsurf-search config set
Exposed tool
web_search
| arg | type | required | notes |
|---|---|---|---|
query |
string | yes | search query |
limit |
number | no | 1–10, default 5 |
domain |
string | no | optional domain filter |
mode |
number | no | optional upstream mode |
Returns MCP text content with JSON:
{ "hits": [ { "title": "...", "url": "...", "snippet": "...", "source": "windsurf" } ] }
Development
node --test test/*.mjs # offline unit + protocol tests
RUN_LIVE_SEARCH=1 npm test # also hit live API if key is configured
Security notes
- Never commit real tokens.
- Session tokens expire; re-run
config setwhen searches return 401. config showonly prints a masked key.- This is not an official Windsurf/Devin product.
License
MIT
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.