MCP.so
Sign In

mcp-ollama

@true-alter

Repository

About mcp-ollama

MCP server wrapping local Ollama models for Claude Code delegation

Config

Add this server to your MCP-compatible client using the configuration below.

{
  "mcpServers": {
    "ollama": {
      "command": "node",
      "args": [
        "/absolute/path/to/mcp-ollama/dist/index.js"
      ],
      "env": {
        "OLLAMA_HOST": "http://localhost:11434",
        "OLLAMA_MODEL": "hermes3:8b"
      }
    }
  }
}

Tools

9

General-purpose generation with system + user prompt

Summarise a blob of text

Analyse text against a specific question

Draft content in a given style

Code tasks: docstring / test / explain / review / types / refactor-suggest

Diff-driven tasks: commit-message / pr-description / changelog / summary / impact

Mechanical code transformations

List models available on the local Ollama host

Pull a model onto the local Ollama host

Overview

What is mcp-ollama?

mcp-ollama is an MCP server that wraps local Ollama models, allowing orchestrators like Claude Code or Cursor to offload simple tasks to a locally‑running model instead of paying per token. It is plumbing: the orchestrator decides what to route locally, and the server exposes nine tools that pass the work straight to a local model.

How to use mcp-ollama?

Install via npm install -g @truealter/mcp-ollama or run with npx @truealter/mcp-ollama. You need a running Ollama instance with at least one model pulled (default hermes3:8b). Configure the server in your MCP client (e.g., Claude Code) using stdio transport and optionally set OLLAMA_HOST and OLLAMA_MODEL environment variables. Tools are exposed automatically via MCP introspection.

Key features of mcp-ollama

  • Nine tools for generation, summarisation, code, diffs, and more
  • Runs entirely over stdio – no network call‑outs by the server
  • Override model per tool call or via OLLAMA_MODEL environment variable
  • Ships no model weights, no telemetry, no analytics
  • Stateless between calls; every request stays on the host running Ollama
  • Supports Docker deployment with automatic host‑network routing

Use cases of mcp-ollama

  • Route classification, one‑liners, and tags to a fast 8B model instead of a frontier API
  • Generate commit messages, PR descriptions, and changelogs locally from diffs
  • Write docstrings, tests, or code reviews using a code‑specialised Ollama model
  • Summarise long documents or analyse text against a specific question without API costs
  • Draft content in a given style, then refine it later with an external model

FAQ from mcp-ollama

What dependencies does mcp-ollama require?

Node 18+, a running Ollama instance (default http://localhost:11434), and at least one model pulled into Ollama.

Does mcp-ollama send my data anywhere?

No. The server makes no network calls beyond the configured OLLAMA_HOST endpoint, and ships no telemetry, analytics, or auto‑update pinger. Data stays on the host where Ollama runs.

What transport does mcp-ollama use?

It uses stdio transport and is designed to be launched by MCP‑aware clients (Claude Code, Cursor, etc.). Running the binary directly is only useful for debugging.

Why are my tool outputs truncated?

Each tool has a default max_tokens of 2048. For longer generations, pass max_tokens explicitly in the tool call.

I get a 404 error when calling a tool.

The model you requested is not available on Ollama. Run ollama pull <name> or use the local_pull tool from your MCP client to pull it.

Frequently asked questions

What dependencies does mcp-ollama require?

Node 18+, a running Ollama instance (default `http://localhost:11434`), and at least one model pulled into Ollama.

Does mcp-ollama send my data anywhere?

No. The server makes no network calls beyond the configured `OLLAMA_HOST` endpoint, and ships no telemetry, analytics, or auto‑update pinger. Data stays on the host where Ollama runs.

What transport does mcp-ollama use?

It uses stdio transport and is designed to be launched by MCP‑aware clients (Claude Code, Cursor, etc.). Running the binary directly is only useful for debugging.

Why are my tool outputs truncated?

Each tool has a default `max_tokens` of 2048. For longer generations, pass `max_tokens` explicitly in the tool call.

I get a 404 error when calling a tool.

The model you requested is not available on Ollama. Run `ollama pull <name>` or use the `local_pull` tool from your MCP client to pull it.

Comments

Basic information

Category

AI & Agents

License

Apache-2.0

Transports

stdio

Author

true-alter

Submitted by

maxchop

More AI & Agents MCP servers

Lumify Sports Intelligence

Lumify

Agent-ready sports intelligence API: live scores, odds, line movement, public betting splits, and explainable bet confidence — via 16 MCP tools at https://lumify.ai/mcp. Get a free key instantly — no signup, email, or ca

Sequential Thinking Multi-Agent System (MAS)

FradSer

An advanced sequential thinking process using a Multi-Agent System (MAS) built with the Agno framework and served via MCP.

306

21st.dev Magic AI Agent

21st-dev

It's like v0 but in your Cursor/WindSurf/Cline. 21st dev Magic MCP server for working with your frontend like Magic

5.3K

bsv.cx

Andy Rowe

Timestamp and verify evidence on-chain. Let your agent prove what it saw and when. bsv.cx gives an AI agent verifiable memory. Timestamp a hash of any output, archive a web page as it looked, or fetch a URL with a recei

Just Prompt - A lightweight MCP server for LLM providers

disler

just-prompt is an MCP server that provides a unified interface to top LLM providers (OpenAI, Anthropic, Google Gemini, Groq, DeepSeek, and Ollama)

733