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

# Use Portal.io in Your AI Tool

> Connect Portal.io's documentation to Claude, ChatGPT, Cursor, and other AI tools using MCP. Get accurate API help directly in the assistant you already use.

The Portal.io documentation is available as a **Model Context Protocol (MCP) server**. If your AI tool supports MCP, you can connect it to Portal.io and ask questions about the API, generate integration code, troubleshoot errors, and look up endpoint details — without leaving your editor or chat.

Your AI tool searches Portal.io's documentation on demand and reads full pages when it needs more detail. The result is responses grounded in the current docs rather than guesses from the model's training data.

## What you can do with it

Once connected, ask your AI tool questions like:

* *"How do I authenticate to the Portal.io API in Python?"*
* *"Generate a webhook subscription that triggers on proposal status change."*
* *"What query parameters does the list-proposals endpoint accept?"*
* *"Walk me through creating a proposal with multiple areas and options."*

The tool will retrieve the relevant pages from docs.portal.io and use them to answer.

<Note>
  This is different from the **AI Builder** endpoints in the API. AI Builder is a Portal.io feature for generating proposals. This page is about connecting *external* AI tools (Claude, ChatGPT, Cursor, etc.) to Portal.io's documentation.
</Note>

## The connection URL

```
https://docs.portal.io/mcp
```

That's all you need. Anonymous access; no authentication required.

## Set up your AI tool

<Tabs>
  <Tab title="Claude Desktop">
    1. Open **Claude Desktop**.
    2. Open **Settings** (⌘`,` on macOS, Ctrl `,` on Windows).
    3. Go to **Connectors**.
    4. Click **Add custom connector**.
    5. Set:
       * **Name**: `Portal.io`
       * **URL**: `https://docs.portal.io/mcp`
    6. Click **Add**.

    Start a new chat and ask a Portal.io API question. Claude will use the connector automatically when relevant.
  </Tab>

  <Tab title="Claude (web)">
    1. Sign in to [claude.ai](https://claude.ai).
    2. Open **Settings** → **Connectors**.
    3. Click **Add custom connector**.
    4. Set:
       * **Name**: `Portal.io`
       * **URL**: `https://docs.portal.io/mcp`
    5. Click **Add**.

    To use the connector in a chat, click the **attachments button** (the plus icon next to the message input), then select **Portal.io** from your connectors list.

    <Note>
      Custom MCP connectors on Claude.ai require a paid plan (Pro, Max, Team, or Enterprise).
    </Note>
  </Tab>

  <Tab title="Cursor">
    **Fastest setup**: open any page on [docs.portal.io](https://docs.portal.io), click the contextual menu at the top of the page, and select **Connect to Cursor**. The connector installs in one click.

    **Manual setup**:

    1. In Cursor, press <kbd>⌘</kbd>+<kbd>Shift</kbd>+<kbd>P</kbd> (<kbd>Ctrl</kbd>+<kbd>Shift</kbd>+<kbd>P</kbd> on Windows) to open the command palette.

    2. Search for **Open MCP settings**.

    3. Click **Add custom MCP**. This opens your `mcp.json` file.

    4. Add:

       ```json theme={null}
       {
         "mcpServers": {
           "Portal.io": {
             "url": "https://docs.portal.io/mcp"
           }
         }
       }
       ```

    5. Save the file.

    In any Cursor chat or composer panel, ask a Portal.io-related question — Cursor will use the connector.
  </Tab>

  <Tab title="VS Code">
    **Fastest setup**: open any page on [docs.portal.io](https://docs.portal.io), click the contextual menu at the top of the page, and select **Connect to VS Code**.

    **Manual setup**: create a `.vscode/mcp.json` file in your workspace and add:

    ```json theme={null}
    {
      "servers": {
        "Portal.io": {
          "type": "http",
          "url": "https://docs.portal.io/mcp"
        }
      }
    }
    ```

    Restart your MCP-aware AI extension (Copilot Chat, Continue, etc.) so it picks up the new server.
  </Tab>

  <Tab title="Claude Code (CLI)">
    If you use [Claude Code](https://docs.claude.com/claude-code), add the connector with one command:

    ```bash theme={null}
    claude mcp add --transport http portal-io https://docs.portal.io/mcp
    ```

    Verify it's connected:

    ```bash theme={null}
    claude mcp list
    ```

    You should see `portal-io` in the list. Ask Portal.io questions in any Claude Code session and the docs will be available as a tool.
  </Tab>

  <Tab title="ChatGPT">
    ChatGPT's MCP support is available through **Custom Connectors** on ChatGPT Business, Enterprise, and Edu plans:

    1. Open **ChatGPT Settings**.
    2. Go to **Connectors** → **Add custom connector**.
    3. Set:
       * **Name**: `Portal.io`
       * **URL**: `https://docs.portal.io/mcp`
    4. Click **Add**.

    For ChatGPT Plus and Free, native MCP support is rolling out gradually. As an alternative, you can connect Portal.io to Claude Desktop or Cursor and use those tools for Portal.io-specific work.
  </Tab>

  <Tab title="Windsurf">
    1. Open **Windsurf**.
    2. Open **Cascade** → **MCP Servers**.
    3. Click **Add custom server**.
    4. Set:
       * **Name**: `portal-io`
       * **URL**: `https://docs.portal.io/mcp`
    5. Save.
  </Tab>
</Tabs>

## Try it out

Once connected, try these prompts to confirm the connector is working and to see what it can do:

<AccordionGroup>
  <Accordion title="Authentication walkthrough">
    *"How do I authenticate to the Portal.io API? Walk me through getting credentials and signing a request."*

    Your AI tool should retrieve the authentication overview and signing-requests guide, then explain HMAC signing with example code.
  </Accordion>

  <Accordion title="Generate webhook code">
    *"Write a Node.js Express endpoint that receives a Portal.io webhook for proposal status changes, validates the signature, and logs the payload."*

    Your AI tool should pull the webhook documentation and event payload schemas, then generate working code.
  </Accordion>

  <Accordion title="Look up an endpoint">
    *"What query parameters does GET /public/proposals support, and what does the response look like?"*

    Your AI tool should retrieve the list-proposals reference and summarize the parameters, defaults, and response shape.
  </Accordion>

  <Accordion title="Plan a multi-step integration">
    *"I want to build a workflow that creates a proposal, adds two areas with options, and assigns it to a contact. Outline the API calls in order with example payloads."*

    Your AI tool should pull the relevant endpoint references and concept pages to plan the sequence.
  </Accordion>
</AccordionGroup>

## Troubleshooting

<AccordionGroup>
  <Accordion title="The tool isn't using the Portal.io connector">
    Some AI tools require you to explicitly invoke the connector. Try prefixing your prompt with the connector name (`@Portal.io`, `@portal-io`) or asking a question that's clearly about Portal.io.

    Also confirm the connector was actually added — open your tool's connector list and verify Portal.io appears.
  </Accordion>

  <Accordion title="The connection fails to add">
    Check the URL is `https://docs.portal.io/mcp` exactly — no trailing slash, no extra path segments.

    If your tool requires you to choose a transport, select `http` (not `stdio`).
  </Accordion>

  <Accordion title="Responses cite outdated information">
    The MCP endpoint is auto-generated from the live documentation at docs.portal.io. If you see incorrect information, the underlying docs likely need updating — please [file an issue](https://portal.canny.io/developer-api) so we can fix the source.
  </Accordion>
</AccordionGroup>

## Feedback

This integration is new — your feedback shapes what we improve. If your AI tool isn't listed here, or you'd like a better experience with a particular tool, [let us know](https://portal.canny.io/developer-api).