Maguyva
Code intelligence for coding agents: semantic, AST, graph, and full-text search. 279+ languages.
Links
README
From the repo.
@maguyva/mcp
Maguyva is a code-intelligence MCP server: semantic, structural, graph, and text search across your indexed repositories, so your agent finds exactly what it needs, fast.
Maguyva is a remote streamable-HTTP MCP server at https://maguyva.tools/mcp. Most modern clients connect to it directly — no local process needed. This package is a stdio bridge for the clients that can't.
Pick your path:
| Your client | Use |
|---|---|
| Claude Code or OpenAI Codex | The maguyva/agent-plugins plugin — bundles the connection, the usage skill, and session-start guidance |
Supports remote MCP with an Authorization header (Cursor, VS Code, Windsurf, Zed, OpenCode, Gemini CLI, …) | Direct HTTP config — see below |
| stdio-only MCP client | This package |
Whichever path you take, get your API key from the maguyva.ai dashboard, and install the usage skill (see Install the skill) — the skill is what teaches your agent when and how to reach for Maguyva's tools.
Claude Code / Codex
Install the plugin from maguyva/agent-plugins:
/plugin marketplace add maguyva/agent-plugins
/plugin install maguyva@maguyva
export MAGUYVA_API_KEY='<your key>'
The plugin connects over HTTP directly and includes the skill and hooks — nothing else to set up.
Direct HTTP (preferred for everything else)
If your client supports remote MCP servers with a custom Authorization header, connect directly:
{
"mcpServers": {
"maguyva": {
"url": "https://maguyva.tools/mcp",
"headers": {
"Authorization": "Bearer ${MAGUYVA_API_KEY}"
}
}
}
}
Exact config shape varies by client (VS Code uses servers + type: "http", Gemini CLI uses httpUrl, Codex uses TOML, …). The installation docs have a verified copy-pasteable snippet for each client.
This package: the stdio bridge
Only needed when your client speaks stdio but cannot connect to a remote HTTP server directly. It runs locally and forwards between your client's stdio transport and https://maguyva.tools/mcp, reconnecting to the remote without dropping the local pipe.
Requires Node.js. Add to your client's MCP config:
{
"mcpServers": {
"maguyva": {
"command": "npx",
"args": ["-y", "@maguyva/mcp"],
"env": {
"MAGUYVA_API_KEY": "your-api-key"
}
}
}
}
Options: pass --api-key=KEY instead of the env var, and --server=URL (or MAGUYVA_SERVER_URL) to override the endpoint.
Install the skill
Whatever the transport, install the Maguyva skill so your agent knows how to use the tools well — tool selection, when local grep beats indexed search, and troubleshooting:
curl -fsSL https://maguyva.tools/install-skill.sh | sh -s -- --client <client>
--client accepts claude, opencode, codex, cursor, vscode, gemini, cline, or agents (the generic ~/.agents/skills path). The skill is authored to the open Agent Skills standard and lives in maguyva/agent-plugins. Claude Code and Codex plugin users already have it — the plugin bundles it.
License
MIT
Config for your environment
Use the endpoint URL below in your config. No API key — you connect directly.
Tool
OS
Config file: ~/.cursor/mcp.json
{
"mcpServers": {
"mcp-server": {
"url": "https://maguyva.tools/mcp"
}
}
}Paste into mcpServers in the config file. Restart Cursor after saving.
If this MCP is also published on mcpchannel.ai, you can subscribe from Browse and use the gateway config there instead.