Developer Hub & Agent Integration

PolyTranscript API & MCP Documentation

Learn how to integrate multi-platform transcription, AI chaptering, and soundbite search into your Python apps, TypeScript web apps, and Claude / Cursor AI agents.

Model Context Protocol (MCP) Integration

Connect to Claude Desktop & Cursor IDE

PolyTranscript includes a native Model Context Protocol (MCP) server. When configured, AI agents can automatically invoke tools to transcribe YouTube, TikTok, and podcasts, generate AI chapters, search soundbites, and cite timestamps during chat conversations.

Claude Desktop Config (claude_desktop_config.json):

{
  "mcpServers": {
    "polytranscript": {
      "command": "python",
      "args": [
        "/path/to/polytranscript/backend/cli.py",
        "mcp"
      ]
    }
  }
}

poly_transcribeTranscribe YouTube/TikTok/Podcast to clean markdown

poly_get_chaptersGenerate timestamped chapters & bullet points

poly_search_soundbitesFind exact quotes and moments with timestamps

REST API Reference

1. Transcribe Media (cURL)

curl -X POST https://api.polytranscript.dev/api/v1/transcribe \
  -H "Content-Type: application/json" \
  -H "X-API-Key: poly_starter_yourkey" \
  -d '{
    "url": "https://www.youtube.com/watch?v=aircAruvnKk",
    "language": "en",
    "include_chapters": true,
    "include_summary": true
  }'

2. Python SDK Integration

import httpx

response = httpx.post(
    "https://api.polytranscript.dev/api/v1/transcribe",
    headers={"X-API-Key": "poly_starter_yourkey"},
    json={
        "url": "https://www.tiktok.com/@user/video/1234567890",
        "include_chapters": True,
        "include_summary": True
    },
    timeout=60.0
)

data = response.json()
print("Title:", data["metadata"]["title"])
print("Summary:", data["summary"]["tldr"])
for chapter in data["chapters"]:
    print(f"[{chapter['formatted_start']}] {chapter['title']}")