Local AI MCP

An operations-first control plane for local LLM runtimes like Ollama, LM Studio, and llama.cpp. It enables agents to discover installed models, check VRAM fit, manage model lifecycle (pull, load, unload), and delegate inference tasks to the best local model without sending data to the cloud.

Package and access

Local package with source-attributed installation metadata. Koskamo does not execute packages during review.

Community listingLocal stdioSource reviewed; not executedPublished by TMHSDigital

You need:Install at least one runtime ( Ollama , LM Studio , llama.cpp , Docker Model Runner, MLX, vLLM or Foundry Local)

Vendor documentation

What you can do with it

Discover installed models and check VRAM fit
Pull, load, and unload models across runtimes
Delegate inference tasks to the best local model
Benchmark model performance and search for fitting GGUFs

Source and maintenance

Community listingRegistry 0.6.1

mcp-registry: io.github.TMHSDigital/local-ai-mcp@0.6.1 Source repository Project website

Maintenance checked: Nov 6, 2026, 10:11 PM UTC

Packages and installation

Source-provided package data. Koskamo has not executed or live-probed this package.

@tmhs/local-ai-mcp 0.6.1

npmnpx

@tmhs/local-ai-mcp package definition

{
  "version": "0.6.1",
  "transport": "stdio",
  "identifier": "@tmhs/local-ai-mcp",
  "runtimeHint": "npx",
  "registryType": "npm",
  "registryBaseUrl": "https://registry.npmjs.org",
  "environmentVariables": [
    {
      "name": "OLLAMA_HOST",
      "default": "http://localhost:11434",
      "description": "Ollama base URL"
    },
    {
      "name": "LMSTUDIO_HOST",
      "default": "http://localhost:1234",
      "description": "LM Studio server base URL"
    },
    {
      "name": "LLAMACPP_HOST",
      "default": "http://localhost:8080",
      "description": "llama.cpp server (llama-server) base URL"
    },
    {
      "name": "DOCKER_MODEL_RUNNER_HOST",
      "default": "http://localhost:12434",
      "description": "Docker Model Runner host TCP endpoint"
    },
    {
      "name": "MLX_HOST",
      "description": "mlx_lm.server base URL (opt-in; Apple Silicon)"
    },
    {
      "name": "VLLM_HOST",
      "description": "vLLM OpenAI server base URL (opt-in)"
    },
    {
      "name": "VLLM_API_KEY",
      "description": "Bearer token for VLLM_HOST (when vLLM runs with --api-key)"
    },
    {
      "name": "FOUNDRY_LOCAL_HOST",
      "description": "Foundry Local service URL (opt-in; port from `foundry service status`)"
    },
    {
      "name": "OPENAI_COMPAT_HOST",
      "description": "Any OpenAI-compatible /v1 server (opt-in)"
    },
    {
      "name": "OPENAI_COMPAT_API_KEY",
      "description": "Bearer token for OPENAI_COMPAT_HOST"
    },
    {
      "name": "MOONSHOT_API_KEY",
      "description": "Moonshot AI (Kimi) key. Hosted: only used when named explicitly."
    },
    {
      "name": "LOCAL_AI_ALLOW_HOSTED_FALLBACK",
      "description": "Allow implicit routing to hosted providers (default: off)"
    },
    {
      "name": "MOONSHOT_HOST",
      "default": "https://api.moonshot.ai/v1",
      "description": "Moonshot API base URL"
    },
    {
      "name": "LOCAL_AI_DISABLE",
      "description": "Comma-separated provider names to skip probing"
    },
    {
      "name": "LOCAL_AI_PROVIDERS",
      "description": "JSON array of extra provider instances: {type, host, name?, apiKey?, kind?}"
    },
    {
      "name": "LOCAL_AI_DISABLE_NETWORK_SEARCH",
      "description": "Keep search_models offline (catalog only)"
    },
    {
      "name": "LOCAL_AI_BENCHMARK_CACHE",
      "description": "Benchmark results file used to rank delegate by measured speed (default: per-user cache dir; off disables)"
    }
  ]
}

Tools

Reported by the attributed source; not independently observed by Koskamo.

No tool definitions available.

Connect from your client

Each snippet configures the same source-provided package; replace placeholders before use.

This server is operated by the named third party. Listing and endpoint checks are not an endorsement or a security audit. Review the operator's permissions, terms, and data handling before connecting. How listings are checked and corrected.

Cursor

Merge this local stdio configuration, replace every placeholder environment value, and restart Cursor. Koskamo has not executed this package.

Cursor configuration

{
  "mcpServers": {
    "local-ai-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@tmhs/local-ai-mcp@0.6.1"
      ],
      "env": {
        "OLLAMA_HOST": "<YOUR_OLLAMA_HOST>",
        "LMSTUDIO_HOST": "<YOUR_LMSTUDIO_HOST>",
        "LLAMACPP_HOST": "<YOUR_LLAMACPP_HOST>",
        "DOCKER_MODEL_RUNNER_HOST": "<YOUR_DOCKER_MODEL_RUNNER_HOST>",
        "MLX_HOST": "<YOUR_MLX_HOST>",
        "VLLM_HOST": "<YOUR_VLLM_HOST>",
        "VLLM_API_KEY": "<YOUR_VLLM_API_KEY>",
        "FOUNDRY_LOCAL_HOST": "<YOUR_FOUNDRY_LOCAL_HOST>",
        "OPENAI_COMPAT_HOST": "<YOUR_OPENAI_COMPAT_HOST>",
        "OPENAI_COMPAT_API_KEY": "<YOUR_OPENAI_COMPAT_API_KEY>",
        "MOONSHOT_API_KEY": "<YOUR_MOONSHOT_API_KEY>",
        "LOCAL_AI_ALLOW_HOSTED_FALLBACK": "<YOUR_LOCAL_AI_ALLOW_HOSTED_FALLBACK>",
        "MOONSHOT_HOST": "<YOUR_MOONSHOT_HOST>",
        "LOCAL_AI_DISABLE": "<YOUR_LOCAL_AI_DISABLE>",
        "LOCAL_AI_PROVIDERS": "<YOUR_LOCAL_AI_PROVIDERS>",
        "LOCAL_AI_DISABLE_NETWORK_SEARCH": "<YOUR_LOCAL_AI_DISABLE_NETWORK_SEARCH>",
        "LOCAL_AI_BENCHMARK_CACHE": "<YOUR_LOCAL_AI_BENCHMARK_CACHE>"
      }
    }
  }
}

Cline

Merge this local stdio configuration, replace every placeholder environment value, and restart Cline. Koskamo has not executed this package.

Cline configuration

{
  "mcpServers": {
    "local-ai-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@tmhs/local-ai-mcp@0.6.1"
      ],
      "env": {
        "OLLAMA_HOST": "<YOUR_OLLAMA_HOST>",
        "LMSTUDIO_HOST": "<YOUR_LMSTUDIO_HOST>",
        "LLAMACPP_HOST": "<YOUR_LLAMACPP_HOST>",
        "DOCKER_MODEL_RUNNER_HOST": "<YOUR_DOCKER_MODEL_RUNNER_HOST>",
        "MLX_HOST": "<YOUR_MLX_HOST>",
        "VLLM_HOST": "<YOUR_VLLM_HOST>",
        "VLLM_API_KEY": "<YOUR_VLLM_API_KEY>",
        "FOUNDRY_LOCAL_HOST": "<YOUR_FOUNDRY_LOCAL_HOST>",
        "OPENAI_COMPAT_HOST": "<YOUR_OPENAI_COMPAT_HOST>",
        "OPENAI_COMPAT_API_KEY": "<YOUR_OPENAI_COMPAT_API_KEY>",
        "MOONSHOT_API_KEY": "<YOUR_MOONSHOT_API_KEY>",
        "LOCAL_AI_ALLOW_HOSTED_FALLBACK": "<YOUR_LOCAL_AI_ALLOW_HOSTED_FALLBACK>",
        "MOONSHOT_HOST": "<YOUR_MOONSHOT_HOST>",
        "LOCAL_AI_DISABLE": "<YOUR_LOCAL_AI_DISABLE>",
        "LOCAL_AI_PROVIDERS": "<YOUR_LOCAL_AI_PROVIDERS>",
        "LOCAL_AI_DISABLE_NETWORK_SEARCH": "<YOUR_LOCAL_AI_DISABLE_NETWORK_SEARCH>",
        "LOCAL_AI_BENCHMARK_CACHE": "<YOUR_LOCAL_AI_BENCHMARK_CACHE>"
      }
    }
  }
}

VS Code

Merge this local stdio configuration, replace every placeholder environment value, and restart VS Code. Koskamo has not executed this package.

VS Code configuration

{
  "servers": {
    "local-ai-mcp": {
      "type": "stdio",
      "command": "npx",
      "args": [
        "-y",
        "@tmhs/local-ai-mcp@0.6.1"
      ],
      "env": {
        "OLLAMA_HOST": "<YOUR_OLLAMA_HOST>",
        "LMSTUDIO_HOST": "<YOUR_LMSTUDIO_HOST>",
        "LLAMACPP_HOST": "<YOUR_LLAMACPP_HOST>",
        "DOCKER_MODEL_RUNNER_HOST": "<YOUR_DOCKER_MODEL_RUNNER_HOST>",
        "MLX_HOST": "<YOUR_MLX_HOST>",
        "VLLM_HOST": "<YOUR_VLLM_HOST>",
        "VLLM_API_KEY": "<YOUR_VLLM_API_KEY>",
        "FOUNDRY_LOCAL_HOST": "<YOUR_FOUNDRY_LOCAL_HOST>",
        "OPENAI_COMPAT_HOST": "<YOUR_OPENAI_COMPAT_HOST>",
        "OPENAI_COMPAT_API_KEY": "<YOUR_OPENAI_COMPAT_API_KEY>",
        "MOONSHOT_API_KEY": "<YOUR_MOONSHOT_API_KEY>",
        "LOCAL_AI_ALLOW_HOSTED_FALLBACK": "<YOUR_LOCAL_AI_ALLOW_HOSTED_FALLBACK>",
        "MOONSHOT_HOST": "<YOUR_MOONSHOT_HOST>",
        "LOCAL_AI_DISABLE": "<YOUR_LOCAL_AI_DISABLE>",
        "LOCAL_AI_PROVIDERS": "<YOUR_LOCAL_AI_PROVIDERS>",
        "LOCAL_AI_DISABLE_NETWORK_SEARCH": "<YOUR_LOCAL_AI_DISABLE_NETWORK_SEARCH>",
        "LOCAL_AI_BENCHMARK_CACHE": "<YOUR_LOCAL_AI_BENCHMARK_CACHE>"
      }
    }
  }
}

Claude Desktop

Merge this local stdio configuration, replace every placeholder environment value, and restart Claude Desktop. Koskamo has not executed this package.

Claude Desktop configuration

{
  "mcpServers": {
    "local-ai-mcp": {
      "command": "npx",
      "args": [
        "-y",
        "@tmhs/local-ai-mcp@0.6.1"
      ],
      "env": {
        "OLLAMA_HOST": "<YOUR_OLLAMA_HOST>",
        "LMSTUDIO_HOST": "<YOUR_LMSTUDIO_HOST>",
        "LLAMACPP_HOST": "<YOUR_LLAMACPP_HOST>",
        "DOCKER_MODEL_RUNNER_HOST": "<YOUR_DOCKER_MODEL_RUNNER_HOST>",
        "MLX_HOST": "<YOUR_MLX_HOST>",
        "VLLM_HOST": "<YOUR_VLLM_HOST>",
        "VLLM_API_KEY": "<YOUR_VLLM_API_KEY>",
        "FOUNDRY_LOCAL_HOST": "<YOUR_FOUNDRY_LOCAL_HOST>",
        "OPENAI_COMPAT_HOST": "<YOUR_OPENAI_COMPAT_HOST>",
        "OPENAI_COMPAT_API_KEY": "<YOUR_OPENAI_COMPAT_API_KEY>",
        "MOONSHOT_API_KEY": "<YOUR_MOONSHOT_API_KEY>",
        "LOCAL_AI_ALLOW_HOSTED_FALLBACK": "<YOUR_LOCAL_AI_ALLOW_HOSTED_FALLBACK>",
        "MOONSHOT_HOST": "<YOUR_MOONSHOT_HOST>",
        "LOCAL_AI_DISABLE": "<YOUR_LOCAL_AI_DISABLE>",
        "LOCAL_AI_PROVIDERS": "<YOUR_LOCAL_AI_PROVIDERS>",
        "LOCAL_AI_DISABLE_NETWORK_SEARCH": "<YOUR_LOCAL_AI_DISABLE_NETWORK_SEARCH>",
        "LOCAL_AI_BENCHMARK_CACHE": "<YOUR_LOCAL_AI_BENCHMARK_CACHE>"
      }
    }
  }
}

Let your agent find its tools

Find servers like Local AI MCP from your agent. Add Koskamo once to search the index for endpoints, access requirements, and client setup. Public discovery needs no key.

Connect Koskamo Explore workflows

See all 5,639