Hugging Face MCP server

What the model can do

Hugging Face Hub: models, datasets and Spaces by author or search with downloads, likes, tags and last change, a repo's discussions and pull requests with their comments and status, a repo's commits, and Jobs with their stage and hardware; when you ask, can open discussions or pull requests, comment, close, reopen or rename them, merge pull requests, create model or dataset repos, start or cancel Jobs on paid hardware, change a Space's hardware, and accept or reject access requests to gated repos.

Reads

• huggingface_list_reposModels, datasets or Spaces on the Hugging Face Hub (type), by author (a user or organization) and/or search text: id, author, private or gated…
• huggingface_list_discussionsDiscussions and pull requests on one Hugging Face repo, newest first by default: number, title, status (open, closed, merged, draft)…
• huggingface_get_discussionOne discussion or pull request on a Hugging Face repo with its timeline: comments (author, time, text; hidden ones say so), status and title changes…
• huggingface_list_commitsCommits on a Hugging Face repo's branch or revision (default main): id, title, message, authors and date, as the Hub returns them.
• huggingface_list_jobsHugging Face Jobs (compute runs) of a namespace: id, stage (RUNNING, SCHEDULING, COMPLETED, ERROR, CANCELED), status message, hardware flavor…
• huggingface_get_jobOne Hugging Face Job: stage and status message, hardware flavor, image or Space, timings and labels.

Acts, when you ask

These change something in Hugging Face. The model is told to use them only when you ask for exactly that, and each one is granted by name: per team on a shared connection, by the owner on a personal one.

• huggingface_create_discussionOpens a discussion on a Hugging Face repo, or a draft pull request when pull_request is true, with a title and description…
• huggingface_comment_on_discussionPosts a comment on a Hugging Face discussion or pull request as the token's user…
• huggingface_set_discussion_statusCloses or reopens a Hugging Face discussion or pull request, optionally with a closing comment; participants are notified.
• huggingface_rename_discussionChanges the title of a Hugging Face discussion or pull request (3 to 200 characters).
• huggingface_merge_pull_requestMerges a pull request into the repo's main branch on the Hugging Face Hub, optionally with a comment.
• huggingface_create_repoCreates an empty model or dataset repo on the Hugging Face Hub under the token's user, or under organization when given.
• huggingface_start_jobStarts a Hugging Face Job: runs the given command in a Docker image (or a Space's image) on the chosen hardware flavor…
• huggingface_cancel_jobCancels a running or scheduled Hugging Face Job; it stops for good (billing stops with it) and cannot be resumed.
• huggingface_set_space_hardwareChanges the hardware a Hugging Face Space runs on (flavor); the Space restarts…
• huggingface_handle_access_requestDecides a person's request to access a gated model or dataset: accepted grants them access, rejected refuses it (with an optional reason they see)…

Connect it

First connect Hugging Face on Dashboard → Connections, then add Memtro to your client once. The Hugging Face tools appear alongside the tools of everything else you have connected; nothing is configured per connector.

claude plugin marketplace add memtro-org/claude-memtro
claude plugin install memtro@memtro --config api_key=sk-memtro-...

{
  "mcpServers": {
    "memtro": {
      "url": "https://platform.memtro.com/api/mcp",
      "headers": {
        "Authorization": "Bearer sk-memtro-..."
      }
    }
  }
}

{
  "mcp": {
    "memtro": {
      "type": "remote",
      "url": "https://platform.memtro.com/api/mcp",
      "headers": {
        "Authorization": "Bearer sk-memtro-..."
      }
    }
  }
}

Who may use it

The key identifies the person, and the server offers that person the tools they may use everywhere else in Memtro. A shared Hugging Face connection is used by the teams an administrator granted it to, tool by tool, on the Permissions page; a personal connection is used by its owner alone, and its tools that act stay off until the owner allows them. See Teams and permissions and Connectors.

Ask it like this

• › Which pull requests are open on our model?
• › Which of our datasets were downloaded most this month?
• › What changed in the last five commits on main?
• › Which Jobs failed in our namespace this week?
• › Comment on discussion 14 that the fix is in the next release.

Questions

Is this an official Hugging Face MCP server?

No. It is Memtro's MCP server with Memtro's Hugging Face connector behind it: the connector reads Hugging Face through its API with the credentials you connect on Dashboard → Connections, and Hugging Face publishes nothing here. The same endpoint carries every other system you connect and Memtro's memory tools.

What does it cost?

Nothing on top of Memtro. One price per organisation by its number of users, from $50 a month for up to 10 users after a 7-day free trial, with every connector, the MCP server, jobs and memory included. You bring your own model keys.

Can it change anything in Hugging Face?

Only when you ask, and only what it has been granted: open a discussion or pull request on a repo; comment on a discussion or pull request; close or reopen a discussion or pull request; rename a discussion or pull request; merge a pull request; create a model or dataset repo; start a Job on paid hardware; cancel a Job; change a Space's hardware; accept or reject a person's access request to a gated repo. Every other tool reads, searches or lists. An administrator grants each acting tool by name per team, and a personal connection acts only after its owner allows it.

Which clients work?

Any MCP client that speaks Streamable HTTP: Claude Code, Cursor, Claude Desktop, OpenCode, Codex, Zed, Windsurf, Cline and the Anthropic and OpenAI SDKs. There is no local process to run; the server is remote and authenticated with your key.

Other MCP servers · Engineering and operations

• GitHub MCP server
• AWS CloudWatch MCP server
• Sentry MCP server
• ServiceNow MCP server
• Okta MCP server
• Twilio MCP server