Advertise
MCP.so
Sign In

Ollama MCP Server

@NewAITees

Repository

Basic information

Category

AI & Agents

Transports

stdio

Author

NewAITees

About Ollama MCP Server

No overview available yet

#ollama#mcp#server

Ollama MCP Server config

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

{
  "mcpServers": {
    "ollama-MCP-server": {
      "command": "python",
      "args": [
        "-m",
        "unittest",
        "discover"
      ]
    }
  }
}

Ollama MCP Server tools

Tool list not available yet

This listing does not currently have a tool list to display. This does not mean the server has no tools; check its documentation for supported capabilities.

We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.

Overview

What is ollama-MCP-server?

ollama-MCP-server is a Model Context Protocol (MCP) server that communicates with a local Ollama LLM instance, providing task decomposition, result evaluation, model management, and workflow management for MCP-compatible applications.

How to use ollama-MCP-server?

Install the package with pip install ollama-mcp-server, then configure the server in your MCP client (e.g., Claude Desktop) using the claude_desktop_config.json file. Tools such as add-task, decompose-task, evaluate-result, and run-model are invoked via the MCP protocol.

Key features of ollama-MCP-server

  • Complex task decomposition into manageable subtasks
  • Evaluation and validation of results against criteria
  • Management and execution of Ollama models
  • Standardized MCP protocol communication
  • Enhanced error handling with structured messages
  • Performance optimization (connection pooling, LRU cache)

Use cases of ollama-MCP-server

  • Break down a large problem into smaller, prioritized subtasks
  • Analyze and score a result based on accuracy, completeness, and clarity
  • Run an Ollama model directly with custom parameters (model, prompt, temperature)
  • Automate multi-step LLM workflows with error recovery

Frequently asked questions

1

What runtime and dependencies are required?

The server requires Python, the `ollama-mcp-server` package, and a local installation of Ollama with at least one model pulled (e.g., `llama3`). Environment variables like `OLLAMA_HOST`, `DEFAULT_MODEL`, and `LOG_LEVEL` can be set.

2

How is the Ollama model selected?

Model selection follows a priority order: (1) the `model` parameter in a tool call, (2) the `env` section of the MCP configuration file, (3) the `OLLAMA_DEFAULT_MODEL` environment variable, and (4) the default value (`llama3`).

3

What performance optimizations are included?

The server uses connection pooling (max 10 concurrent connections per host) and an LRU cache (max 100 entries) to reduce response times and lower Ollama server load. These settings are adjustable in `config.py`.

4

How are errors reported?

Errors return a structured JSON response with a `message`, `status_code`, and optional `details` (e.g., the provided ID for a missing task). If a model is not found, the error message includes the list of available models.

5

What transports and authentication are supported?

The server communicates over stdio using the MCP protocol. The README does not mention authentication mechanisms; access depends on local Ollama and MCP client configuration.

Comments

More AI & Agents MCP servers

L

Legion MCP

faulkj

MCP-native LLM councils for debates, juries, blind panels, voting, refinement, and custom multi-model deliberation.

1

SlopIt: blogs for AI agents

SimbaStack

Blog MCP server for AI agents. One signup call returns a blog, an API key and a live URL; then your agent publishes markdown posts and gets public URLs back. Your agent does the writing; you get a dashboard for stats.

CrawlForge MCP Server

mysleekdesigns

28-tool MCP server for web scraping, crawling, deep research & autonomous extraction — clean Markdown & structured JSON for Claude, Cursor & any MCP client. 1,000 free one-time credits, local-Ollama LLM support.

2

Pocket Network Agentic Marketplace: Pay-per-Call APIs

Pocket Network Foundation

MCP server for the Pocket Network agentic marketplace: search our catalog of data services and utilities and call them, paying per request in USDC on Base with x402. No account or API key. Search and describe are free; p

TruVerifAI Panel-review

TruVerifAI

TruVerifAI's Panel Review is the guardian agent for AI's highest-stakes coding decisions: four frontier models argue over an AI coding agent's riskiest designs, diffs, and commits before they execute, with local review g

1
Advertise