MCP CheckMCP servers for Ollama
21 ways to connect Ollama to Claude, Cursor or VS Code. 2 are hosted, so there's nothing to install; the rest run on your machine. Each row shows the security checks, the secrets asked for and an install command.
Ollama MCP servers 21
Now showing are unlisted. These entries didn't pass our listing filter: no source repo, a very short or duplicated description, or a publisher with 100+ servers. Same automated checks, shown as-is. How we check
| Type | Secrets | Version | Actions | ||||
|---|---|---|---|---|---|---|---|
★ 18 io.github.Lykhoyda/ask-ollama Bridge Claude with local Ollama LLMs for private AI-to-AI collaboration — no API keys, fully local | 7/7All passed | 64health of 100 | npmRuns locally | Noneno keys asked | 0.7.0 | ||
io.github.setheerwagen/mcp-ollama-vllm Call your local Ollama or vLLM model over MCP with schema-validated JSON output | 7/7All passed | 56health of 100 | PyPIRuns locally | 1secret asked | 1.0.6 | ||
io.github.bebopp187-stack/titan-frameworks Review and rewrite LangChain, LlamaIndex, Ollama, and XRPL code against current APIs. | 3/3All passed | 48health of 100 | RemoteHosted | Noneno keys asked | 1.0.1 | ||
io.github.true-alter/mcp-ollama Wraps local Ollama models so bulk generation runs on your own hardware, not a paid API. | 7/7All passed | 52health of 100 | npmRuns locally | Noneno keys asked | 0.2.1 | ||
★ 28 io.github.LogneBudo/llmxray What went wrong with your own Ollama calls, in sentences with the fix. Local, read-only. | 7/7All passed | 70health of 100 | npmRuns locally | Noneno keys asked | 0.8.12 | ||
★ 18 io.github.Lykhoyda/ask-llm Unified MCP server for Gemini, Codex, Claude, Grok, Ollama, and Antigravity | 7/7All passed | 64health of 100 | npmRuns locally | 2secrets asked | 1.2.1 | ||
★ 11 io.github.pvnc228/local-coding-agent Delegate small coding tasks to local Ollama/llama.cpp models; patches gated by real test runs. | 7/7All passed | 55health of 100 | PyPIRuns locally | Noneno keys asked | 1.1.1 | ||
★ 6 io.github.iamjrmh/mcpbridge Unified MCP server bridging Claude Code and Ollama to Roblox Studio and Blender. | 7/7All passed | 46health of 100 | npmRuns locally | Noneno keys asked | 1.0.0 | ||
★ 4 io.github.BlackFoil/claude-token-saver-mcp MCP server that offloads routine coding tasks to local LLM (Ollama) to save Claude API tokens | 7/7All passed | 44health of 100 | npmRuns locally | Noneno keys asked | 0.3.2 | ||
★ 2 io.github.JoaoBerne/cli-bridge-mcp Other AI CLIs (Gemini, GPT, Claude, opencode, Ollama) as a council — ban-safe, no API keys. | 6/71 warning | 54health of 100 | PyPIRuns locally | Noneno keys asked | 0.3.0 | ||
★ 2 io.github.MakeaMouse/fish-bridge-mcp Compresses AI chat sessions into a typed knowledge graph. Local/offline via Ollama. | 7/81 warning | 40health of 100 | PyPIRuns locally | 3secrets asked | 0.1.3 | ||
★ 2 io.github.TheAiSingularity/agentic-research Local research agent that verifies its own answers. Runs on Gemma 3 4B + Ollama, $0/query. | 7/7All passed | 40health of 100 | PyPIRuns locally | 1secret asked | 0.1.3 | ||
★ 1 io.github.sezzlee/llm Let a planning agent hand bounded language work to a local Ollama model. | 7/7All passed | 50health of 100 | npmRuns locally | Noneno keys asked | 0.1.3 | ||
★ 1 io.github.zhaoxinghua09-cell/medxpert-library-connector Local RAG MCP server for your own .md knowledge base: BM25 + local Ollama, offline, zero egress. | 2/2All passed | 50health of 100 | RemoteRuns locally | Noneno keys asked | 1.0.0 | ||
★ 1 io.github.DalecB/obsidian-semantic-mcp Read-only semantic search MCP server for Obsidian vaults using local Ollama embeddings. | 7/7All passed | 52health of 100 | npmRuns locally | Noneno keys asked | 0.2.1 | ||
★ 1 io.github.egoughnour/massive-context-mcp Handles 10M+ token contexts with chunking, sub-queries, and local Ollama inference. | 6/71 warning | 36health of 100 | PyPIRuns locally | Noneno keys asked | 3.0.1 | ||
io.github.qso-graph/qsp-client QSP — relay MCP tools to any OpenAI-compatible local LLM (llama.cpp, Ollama, vLLM) | 7/7All passed | 52health of 100 | PyPIRuns locally | Noneno keys asked | 0.5.0 | ||
io.github.LuminariSoftwares/stack-guardian Health doctor for a home AI stack: Ollama nodes, GPUs and tailnet, with graded confidence. | 7/7All passed | 61health of 100 | PyPIRuns locally | Noneno keys asked | 0.4.0 | ||
io.github.gabrielmahia/offline-mcp Local AI over Ollama for low connectivity: status, local inference, model and deployment guides. | 7/7All passed | 57health of 100 | PyPIRuns locally | Noneno keys asked | 0.1.10 | ||
com.dougalpaw/pawpulse Heartbeat and tokens/sec monitoring for self-hosted LLM servers (Ollama etc.), with down alerts. | 4/4All passed | 43health of 100 | RemoteHosted | Noneno keys asked | 0.3.0 | ||
io.github.sammyboi81/algernon Fan a goal out to cheap parallel LLM workers on your own key or a free local Ollama. Keep your mind. | 7/7All passed | 51health of 100 | PyPIRuns locally | Noneno keys asked | 0.1.1 |
Showing 1 to 21 of 21. These are the most recently updated; search or pick a category for the rest.
Ollama MCP servers: questions people ask
What's the most popular MCP server for Ollama?
Among community builds, the most-starred on GitHub is ask-ollama (★ 18). If you want one that passed every check, ask-ollama did.
Is it safe to give an AI a Ollama MCP server?
On registry checks, 18 of 21 come out clean. We also read the code or live tool list of 19 of them without running anything, and 1 raised a warning worth reading, such as an install script or text aimed at the AI. Treat a pass as "nothing found", and give any server the least access that works.
How do I add a Ollama MCP server to Claude?
To connect Ollama through ask-ollama, use this Claude Desktop (npm) entry: {
"mcpServers": {
"ask-ollama": {
"command": "npx",
"args": [
"-y",
"@ask-llm/ollama-mcp@0.7.0"
]
}
}
}. Pin the version shown before you rely on it.
Do I have to install a Ollama MCP server?
2 can run hosted, so you only paste a URL into Claude, Cursor or VS Code. The 19 that run locally ship as 10 PyPI and 8 npm packages.
Which Ollama MCP servers need no API key?
17 of the 21 list no secrets, including ask-ollama, Titan Frameworks (LangChain, LlamaIndex, Ollama, XRPL) and ~alter mcp-ollama. Some still ask you to sign in once they run.
Who makes these Ollama MCP servers?
No company-verified one yet (checked 31 Oct 2026). All 21 Ollama servers are community builds from 20 different publishers, so check who is behind one before you hand it a Ollama key.
More topics: every “MCP server for…” page. Prefer a skill? Claude skills for this (29).