Metrillm MCP Server

Benchmark your local LLM models in one command. Speed, quality, hardware fitness - with a shareable score and public leaderboard.

People who work with ai benchmark and ollama and want it reachable from Claude, Cursor, VS Code, or another MCP client. The project is written in TypeScript.

VERIFIED ACTIVE

LAST COMMIT 2026-06-02 · ★ 8 · #67 OF 106 MAINTAINED AI & ML · VERIFIED 2026-09-18

Apache-2.0 · TypeScript servers · how we verify → /methodology

01 · Install Metrillm

Claude Code

claude mcp add metrillm-metrillm -- npx -y metrillm-mcp

Claude Desktop / Cursor / VS Code - add to config

{
  "mcpServers": {
    "metrillm-metrillm": {
      "command": "npx",
      "args": [
        "-y",
        "metrillm-mcp"
      ]
    }
  }
}

Same JSON for Cursor. For VS Code, rename the top-level key from `mcpServers` to `servers`.

Using another client? Same JSON, different key

Claude Desktop · mcpServers

Cursor · mcpServers

VS Code · servers

Windsurf · mcpServers

Zed · context_servers

Cline · mcpServers

Roo Code · mcpServers

Continue · mcpServers

LibreChat · mcpServers

Gemini CLI · mcpServers

Codex CLI · mcp_servers

Full setup guides: every client.

02 · Evidence

Security posture

What to check before giving this server access to your agent - from the registry, GitHub, and our own probes. We don't score safety; we show what's verifiable.

runs as local process (stdio) - runs on your machine with your user's permissions

license Apache-2.0 - declared in the repository

npm package metrillm-mcp - unscoped; check the name against the project README before installing

registry namespace io.github.MetriLLM is GitHub-verified and matches the repo owner

03 · What Metrillm can do

Prose above is summarized from the project's README and registry record - no invented capabilities.

Latest releases

v0.2.6 · 2026-03-16

Fixed · Preserve legacy auto-share opt-out preference when migrating user settings to new format.

v0.2.5 · 2026-03-14

Fixed · Harden LM Studio request negotiation: extract shared retry loop, cache output limit mode per endpoint, and add explicit error when both max_output_tokens and max_tokens are rejected.

v0.2.4 · 2026-03-14

Fixed · Include LM Studio models from v0 listing API fallback.

04 · Who maintains Metrillm

metrillm is maintained by metrillm. It's the only MCP server we track from this author; the repo dates to Mar 2026.

05 · Facts

repository: github.com/metrillm/metrillm
category: AI & ML - ranked #67 of 106 actively-maintained AI & ML servers as of 2026-09-18.
registry: io.github.MetriLLM/metrillm (active, first published 2026-03-02 · 8 versions)
packages: npm:metrillm-mcp

06 · Metrillm FAQ

What is Metrillm?

Benchmark your local LLM models in one command. Speed, quality, hardware fitness - with a shareable score and public leaderboard.

Is Metrillm still maintained?

Yes - as of 2026-09-18, its last commit was 2026-06-02. We re-verify nightly.

How do I install Metrillm?

Run `npx -y metrillm-mcp`. You can also paste the ready-made client config above.

Does Metrillm run locally?

Yes - it's a stdio server: it runs on your machine (via npx) with your user's permissions. Your data stays local unless the server itself calls external APIs.

07 · Alternatives to Metrillm

Alternatives to Metrillm

Maintained AI & ML servers if Metrillm isn't the fit.

Pairs well with

Servers that cover what Metrillm doesn't - only shown when the pairing reason fits the companion.

More AI & ML MCP servers · Ollama Handoff · Claude Prompts MCP · Grantex · Prompt Protection

More TypeScript MCP servers · Metrxbots MCP Server · Apple Calendar (mgcrea) · Apple Contacts (mgcrea) · Apple Maps (mgcrea) · Apple Messages (mgcrea) · see all