# Hugging Face MCP server > Search Hugging Face Hub models, datasets, Spaces and papers, and browse repo files and trending lists from your assistant. - Listing: https://mcp.tc/i/hugging-face - Connect: use the server's own URL `https://huggingface.co/mcp` (sign-in optional); clients connect to it directly. The listing link is a page, not an MCP endpoint. - Type: remote (Streamable HTTP) - Auth: optional (works without sign-in) - Category: [AI & Machine Learning](https://mcp.tc/c/ai-ml) - Vendor: Hugging Face - Verified: yes, mcp.tc checked that this is the official server (https://mcp.tc/verify). It says who runs the server, not that it is safe. - Homepage: ## About Connects to the Hugging Face Hub so an assistant can search models, datasets, Spaces, collections, papers and docs, inspect repository details and dataset structure, and read files through hf:// paths. It can also list trending models, datasets, Spaces and daily papers. The registry entry also describes access to Gradio applications. Runs as a hosted streamable HTTP endpoint at https://huggingface.co/mcp. It works anonymously with rate limits. Adding ?login starts OAuth sign-in, or you can send an Authorization Bearer header with a Hugging Face token for higher limits. ## What it can do - Search models, datasets and Spaces by query, author, filters and sort - Get details for repos, including dataset structure and previews - List trending models, datasets, Spaces and papers - Read daily papers and paper content - Read files from Hub repos via hf:// paths - Check the current Hugging Face account and organizations ## Tools (4) - `hf_whoami`: Show the current Hugging Face account, organizations and credential access without revealing… (read-only) - `hub_repo_search`: Search models, datasets or Spaces with query, author, filters, sort and limit. (read-only) - `hub_repo_details`: Get details for repos, including dataset structure and previews. (read-only) - `hf_fs`: Run ls, cat and search over hf:// paths for models, datasets, Spaces, papers and docs. (read-only) ## Example prompts - "Show me today's trending models on Hugging Face" - "Find text-to-image Spaces and link to the top results" - "Show the structure and a preview of the squad dataset" - "List the latest daily papers on Hugging Face" ## Install ### Claude Code 1. Run this in a terminal, in your project folder: ```bash claude mcp add --transport http hugging-face https://huggingface.co/mcp ``` 2. Start Claude Code and type `/mcp`. **hugging-face** should show as connected. To sign in for more access, pick it and choose **Authenticate**. Add `--scope user` to make it available in every project, not just this one. Works without a key. An API key from Hugging Face raises the limits: send it as `Authorization: Bearer `. ### Claude Desktop 1. Open **Settings → Connectors** and click **Add custom connector**. 2. Name it **Hugging Face** and paste this URL: ```url https://huggingface.co/mcp ``` 3. Click **Add**, then **Connect**. Signing in to Hugging Face is optional. Claude Desktop’s JSON config file only starts local servers. Remote servers go through Connectors, and connectors you add on claude.ai show up here too. ### claude.ai 1. Open the connector form on claude.ai. This button fills in the name and URL for you: [Add to claude.ai]() (opens connector settings) 2. Check that the URL reads `https://huggingface.co/mcp` and click **Add**. 3. Click **Connect**. Signing in to Hugging Face is optional. Free plans allow one custom connector. On Team and Enterprise plans an owner adds it under **Organization settings → Connectors**. ### ChatGPT 1. On chatgpt.com, open **Settings → Security and login** and turn on **Developer mode**. 2. Go to `chatgpt.com/plugins` and click **+** to create an app for a remote MCP server. 3. Paste `https://huggingface.co/mcp` as the server URL and choose **No authentication**, or **OAuth** to sign in to Hugging Face for more access. Developer mode is available on the web for Plus, Pro, Business, Enterprise and Education accounts. ### Cursor [Add to Cursor]() (opens Cursor) Or add it by hand to `~/.cursor/mcp.json` (all projects) or `.cursor/mcp.json` (this project): `mcp.json`: ```json { "mcpServers": { "hugging-face": { "url": "https://huggingface.co/mcp" } } } ``` Works without a key. An API key from Hugging Face raises the limits: send it as `Authorization: Bearer `. ### VS Code [Install in VS Code]() (opens VS Code) Or from a terminal: ```bash code --add-mcp '{"name":"hugging-face","type":"http","url":"https://huggingface.co/mcp"}' ``` Or commit it to the repo in `.vscode/mcp.json`: `.vscode/mcp.json`: ```json { "servers": { "hugging-face": { "type": "http", "url": "https://huggingface.co/mcp" } } } ``` Works without a key. An API key from Hugging Face raises the limits: send it as `Authorization: Bearer `. ### Devin Desktop 1. Add it to `~/.config/devin/mcp_config.json` (macOS and Linux) or `%APPDATA%\devin\mcp_config.json` (Windows): `mcp_config.json`: ```json { "mcpServers": { "hugging-face": { "serverUrl": "https://huggingface.co/mcp" } } } ``` 2. Refresh the MCP server list in Cascade. Devin Desktop is the new name for Windsurf. It reads `serverUrl` (or `url`) for remote servers. Works without a key. An API key from Hugging Face raises the limits: send it as `Authorization: Bearer `. ### Codex ```bash codex mcp add hugging-face --url https://huggingface.co/mcp ``` Or edit `~/.codex/config.toml` directly: `config.toml`: ```toml [mcp_servers.hugging-face] url = "https://huggingface.co/mcp" ``` Works without a key. An API key from Hugging Face raises the limits: send it as `Authorization: Bearer `. ### Gemini CLI ```bash gemini mcp add --transport http hugging-face https://huggingface.co/mcp ``` Then, inside Gemini CLI, run `/mcp auth hugging-face` to sign in (optional). This adds it to the current project. Add `-s user` to use it everywhere. Works without a key. An API key from Hugging Face raises the limits: send it as `Authorization: Bearer `. ### Any client Most clients accept this shape. Some name the URL field differently: `serverUrl` in Devin Desktop, `httpUrl` in Gemini CLI’s settings file. ```json { "mcpServers": { "hugging-face": { "type": "http", "url": "https://huggingface.co/mcp" } } } ``` Zed puts servers under `context_servers` in its settings. Cline needs `"type": "streamableHttp"`, or it assumes SSE. Client only starts local servers? Bridge it with `npx -y mcp-remote https://huggingface.co/mcp`. Works without a key. An API key from Hugging Face raises the limits: send it as `Authorization: Bearer `. ## Details - Server version: 0.4.27 - MCP protocol version: 2026-07-28 - Last checked: 2026-10-03 (reachable) - Listed: 2026-10-03 - Updated: 2026-10-03 --- Source: https://mcp.tc/i/hugging-face (mcp.tc is an independent directory, not affiliated with this server's publisher). Corrections: https://mcp.tc/report