> ## Documentation Index
> Fetch the complete documentation index at: https://docs.biohub.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# Connect to Biohub MCP and run your first query

> Add the Biohub MCP server to Codex, Claude, ChatGPT, or Cursor in a few minutes, confirm its seven tools, and run a first protein lookup.

This quickstart connects Biohub MCP to your AI client, checks that its seven tools are available, and runs one request with a known answer.
You don't need a Biohub account or an API key.

## Before you begin

You need an MCP client that can connect to a remote Streamable HTTP server, such as Codex, Claude, ChatGPT, or Cursor.
Some clients limit custom servers by plan or by admin setting.
[Connect your client](/biohub-mcp/connect-clients) lists those requirements.

<Steps>
  <Step title="Add Biohub MCP to your client">
    <Tabs>
      <Tab title="Codex">
        ```bash theme={null}
        codex mcp add biohub --url https://biohub.ai/mcp
        ```

        Run `codex mcp list` to check it.
        The auth column reads `Unsupported`, which is expected because the server has no sign-in.
        Start a new `codex` session to use it.

        The same entry works in the ChatGPT desktop app and the Codex IDE extension.
        To add it from the desktop app instead, see [Connect your client](/biohub-mcp/connect-clients#chatgpt-desktop-app-and-codex-ide-extension).
      </Tab>

      <Tab title="Claude Code">
        ```bash theme={null}
        claude mcp add --transport http biohub https://biohub.ai/mcp
        ```

        Run `claude mcp list` to check it.
        Start a new `claude` session to use it, and enter `/mcp` there to see the server.
      </Tab>

      <Tab title="Claude.ai / Claude Desktop">
        1. Open **Customize**, then **Connectors**.
        2. Click **+**, then **Add custom connector**.
        3. Enter `https://biohub.ai/mcp` as the remote MCP server URL and `Biohub` as the name, leave **Advanced settings** empty, and click **Add**.
        4. In a chat, click **+**, choose **Connectors**, and turn on Biohub.
      </Tab>

      <Tab title="ChatGPT">
        Developer mode is available on the web for Plus, Pro, Business, Enterprise, and Education accounts.

        1. In ChatGPT on the web, open **Settings**, then **Security and login**, and turn on **Developer mode**.
        2. Open [chatgpt.com/plugins](https://chatgpt.com/plugins) and select **+** to create a developer-mode app for a remote MCP server.
        3. Enter `Biohub` as the name and `https://biohub.ai/mcp` as the MCP server URL, choose **No Authentication**, and create it.
        4. In a chat, choose **Developer mode** from the **+** menu and select Biohub.
      </Tab>

      <Tab title="Cursor">
        Add this entry to `~/.cursor/mcp.json`, keeping any servers already listed there:

        ```json theme={null}
        {
          "mcpServers": {
            "biohub": {
              "url": "https://biohub.ai/mcp"
            }
          }
        }
        ```

        Check it in Cursor's MCP settings.
      </Tab>
    </Tabs>
  </Step>

  <Step title="Confirm the tools">
    <Prompt description="List the Biohub MCP tools you can use, without calling any of them.">
      List the Biohub MCP tools you can use, without calling any of them.
    </Prompt>

    You should see these seven tools, possibly with a prefix such as `mcp__biohub__`:

    * `esm_atlas_search_uniprot`
    * `esm_atlas_lookup_accession`
    * `esm_atlas_get_protein_details`
    * `esm_atlas_search_similar_protein_clusters`
    * `esm_atlas_get_cluster_info`
    * `esm_atlas_get_sae_feature_detail`
    * `ui_show_protein_structure`

    If tools are missing, see [Troubleshooting](/biohub-mcp/troubleshooting#no-tools-show-up).
  </Step>

  <Step title="Run a first request">
    <Prompt description="Use Biohub to look up UniProt accession P42212. Tell me the protein name, organism, gene, and sequence length. Don't print the full sequence.">
      Use Biohub to look up UniProt accession P42212.
      Tell me the protein name, organism, gene, and sequence length.
      Don't print the full sequence.
    </Prompt>

    The assistant calls `esm_atlas_search_uniprot` and should report Green fluorescent protein (gene `GFP`) from *Aequorea victoria*, 238 residues long.
    That matches the [UniProt record for P42212](https://www.uniprot.org/uniprotkb/P42212).
  </Step>
</Steps>

## Next steps

* Ask for proteins similar to GFP, or view a structure, with the [Example prompts](/biohub-mcp/example-prompts).
* Set up another client with [Connect your client](/biohub-mcp/connect-clients).
* Learn what the tools accept and return in the [ESM Atlas tools](/biohub-mcp/esm-atlas-tools) reference.