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

# Connect with MCP

> Connect an AI client to PagePith’s remote MCP server.

Connect PagePith’s remote MCP server to give an AI client access to public web pages and social video transcripts.

## Connection

Use the Streamable HTTP endpoint and send your PagePith API key as a Bearer token.

```json theme={null}
{
  "url": "https://mcp.pagepith.com/mcp",
  "headers": {
    "Authorization": "Bearer pp_live_your_key"
  }
}
```

The server is stateless. Every request carries its own authentication.

## Available tools

| Tool            | Purpose                                                                                    |
| --------------- | ------------------------------------------------------------------------------------------ |
| `read_url`      | Return the readable Markdown from a public web page or PDF.                                |
| `process_video` | Queue transcript and metadata processing for a public YouTube, Instagram, or TikTok video. |
| `get_video_job` | Poll a video job and return its result when complete.                                      |

### Read a URL

`read_url` requires a complete `url`. Set `fresh` to `true` to bypass the cache.

### Process a video

Call `process_video` with a public video URL. Then pass its `jobId` to `get_video_job` until the job completes.

<Info>
  MCP requests use the same PagePith account, credits, and API keys as REST requests. A scrape is billed once by the API.
</Info>