--- title: "Hugging Face MCP: review, setup & limitations · Undominated.ai" canonical: https://undominated.ai/mcp-servers/hugging-face/ description: "Connects an assistant to Hub discovery, with optional community tools and account-scoped compute or repository actions." --- # Hugging Face MCP: review, setup & limitations · Undominated.ai > Connects an assistant to Hub discovery, with optional community tools and account-scoped compute or repository actions. [← Explore all mcp servers](/mcp-servers/) ML RESOURCE DISCOVERY / Hugging Face # Hugging Face MCP Connects an assistant to Hub discovery, with optional community tools and account-scoped compute or repository actions. See setup guidance ↓Original source ↗ SOURCE REVIEW Reviewed 2026-09-21 Evidence 5 linked sources Publisher Hugging Face Licence Not established Read what was—and wasn’t—checked ↓ “Search and explore Hub resources: models, datasets, Spaces, papers and more.” Hugging Face · upstream description ↗ Our analysis follows below. 01 / THE REASONING ## Why this made the selection. - Combines specialist model, dataset, paper and documentation discovery in an existing Hub workflow. - The MCP settings distinguish extra tools and community Spaces, so users can inspect the requested capabilities before enabling them. ### A good fit for - Finding relevant models, datasets and papers for an ML project. - Using a deliberately selected Space from an existing assistant. ### Weigh up before choosing - Optional tools extend beyond search: repository contributions, Jobs and Sandboxes can write data or use compute. - Community Spaces add their own code, data-handling and availability boundaries; a Hub listing is not an assurance of their behavior. - The inspected hosted-service documentation does not establish its implementation language or a server-source redistribution licence. 02 / THE REVIEW RECORD ## What we actually inspected. Source review has boundaries. A clear record is more useful than a “safe” badge. ### Material inspected - Official MCP guide: current built-in tools, settings and optional capabilities. - Agents overview: Claude Code connection command. - Spaces MCP guide and Jobs billing documentation: token, quota and compute implications. - Official registry entry co.huggingface/hf-mcp-server: remote transport and endpoint. ### Our findings - The current guide names hf_fs for Hub navigation; optional repository, Sandbox and Jobs tools expand account authority. - Space tools can consume the user’s compute quota, while Jobs have separate hardware-usage billing. ### Not established by this review - Hosted implementation, authentication, community tools, compute execution and data retention were not tested. The review applies to the material and revision named here. A newer upstream release can change its behavior. 03 / PUT IT TO WORK ## Connect a server deliberately. Upstream setup instructions ↗ DOCUMENTED COMMAND claude mcp add hf-mcp-server -t http "https://huggingface.co/mcp?login" Copy command ↗ Copying does not execute this command. It may retrieve a newer version than the reviewed source. - For Claude Code use the documented HTTP command; for other clients copy the generated configuration from your Hugging Face MCP settings. - Select the required tools and Spaces deliberately, then reload the client. ### Before you start - A Hugging Face account for personalized setup and the selected MCP client. - Appropriate token/account permissions and compute access for the tools you enable. ### Compatibility MCP clients supported by the account’s generated setup instructions. · The copyable command is specifically the documented Claude Code HTTP setup. ### Transports Streamable HTTP ### Hosted Hub tools / optional compute and writes - Searches Hub resources and calls selected community Space tools. - Optional capabilities create repository content, manage Sandbox files or run and schedule Jobs. ### Cost model Space calls consume applicable compute quota. Jobs are billed for hardware usage; check the selected service and account terms. 04 / FOLLOW THE EVIDENCE ## The source trail. Our notes are separate from the original resource. Check upstream before adopting a new version. - Official MCP capabilities and settings ↗ Checked 2026-09-21 https://huggingface.co/docs/hub/agents-mcp Supports: upstreamDescription, descriptionSourceUrl, summary, whySelected, bestFor, limitations, install, compatibility, review, access.permissions - Client setup instructions ↗ Checked 2026-09-21 https://huggingface.co/docs/hub/agents-overview Supports: install, compatibility - Community Spaces and quotas ↗ Checked 2026-09-21 https://huggingface.co/docs/hub/spaces-mcp-servers Supports: limitations, install.prerequisites, review, access - Jobs billing ↗ Checked 2026-09-21 https://huggingface.co/docs/hub/jobs-pricing Supports: review, access.cost - Official MCP registry listing ↗ Checked 2026-09-21 https://registry.modelcontextprotocol.io/v0.1/servers?search=huggingface&limit=10 Supports: transports KEEP COMPARING ## Other approaches to consider. Related by category or shared topics. These are alternatives to inspect, not a measured quality order. [### Langfuse data MCP ↗ Inspect Langfuse observations, prompts and evaluation data, and manage supported project resources through its native MCP endpoint.](/mcp-servers/langfuse/) [AI Tools ↗](/tools/)[Skills ↗](/skills/)[Agents ↗](/agents/)[MCP Servers ↗](/mcp-servers/)