LM Studio runs open-weight models (Llama, Mistral, Qwen and friends) entirely on your own machine, and since mid-2025 it has built-in MCP support — which means a fully local assistant can use the same tool ecosystem as the cloud agents: read files with your permission, drive a browser, query databases. Nothing leaves your machine except what a server itself sends out, which is exactly why people pair local models with MCP.
One caveat shapes this list: a model can only use tools if it supports tool calling, so pick a model with function-calling support in LM Studio. The servers below all run as local stdio commands — the configuration LM Studio documents — and clear this site's standard bar of 10+ GitHub stars, 100+ monthly downloads or a verified official tier, ranked by stars, recency and downloads.
A data-ranked shortlist of the 12 strongest MCP servers for LM Studio — every entry links to its full page with copy-paste configs. List compiled 2026-10-28; rankings refresh as the underlying data updates.
How entries are picked: An entry is eligible only if it meets at least one of: 10+ GitHub stars, 100+ monthly npm/PyPI downloads, or a verified official/reference tier on its brand page; entries meeting none of the three are excluded, as are archived repositories. Eligible entries are ranked by a score built from: GitHub stars — multiplied by 10 when the entry has a verified official/reference tier, so official status counts as ten times the community traction rather than a flat bonus; maintenance recency (last push); and npm/PyPI downloads where available. Two tie-break rules keep the list honest: when the same repository is shared by several listings, its star count does not contribute to ranking (the shared figure would otherwise swamp independent projects), and when several implementations share a product name, only the highest-scoring one is listed — the others stay in the directory. Data comes from the public MCP Registry (description, install method), GitHub repository stats and package download counts — no editorial scoring, no paid inclusion, and sponsorship never affects these lists.
Connecting MCP servers to LM Studio
LM Studio manages MCP servers from its Integrations tab, which edits the app's mcp.json: each entry is a command definition (npx, uvx or docker) plus optional environment variables — the same stdio shape the other clients use, so the JSON on each entry's full page maps over directly. If you do not see an Integrations or MCP section, update LM Studio; support landed in the 0.3.6 line in 2025. After adding a server, start it from the same tab and confirm your loaded model supports tool use before assuming the tools are broken. Remote endpoints: LM Studio's mcp.json is command-oriented, so treat remote URL support as version-dependent and check the app's current documentation.
Built-in MCP support arrived in the 0.3.6 release line in 2025, managed from the Integrations tab where the app edits its mcp.json. If your build has no such tab, update the app first — older versions cannot load MCP servers at all.
How do I add an MCP server in LM Studio?
Open the Integrations tab, add a server entry and give it the command from the entry's full page on this site (npx, uvx or docker plus args), along with any environment variables it needs. Start the server there; it then appears as a tool source for capable loaded models.
Why don't tools work with my local model in LM Studio?
Tool calling is a model capability, not an LM Studio setting: the loaded model must support function calling for MCP tools to be offered. Models without tool support will simply ignore the servers. Check the model's card for tool-use support, then confirm the server itself shows as running in the Integrations tab.
Can LM Studio connect to remote MCP servers?
LM Studio's documented mcp.json format is command/stdio-oriented; remote URL support depends on your app version. Everything on this page runs locally as a command, so none of it depends on that answer — pick a remote-capable client like Claude Code or Cursor if remote endpoints are the requirement.