OpenClaw

Overview#

OpenClaw is an open source framework for autonomous AI agents that run on your own machine. By default, web scrapers and local browser automation get blocked by platform bot protections or encounter strict request limits.

Connecting TranscriptDock to OpenClaw gives your autonomous digital employees direct access to YouTube and TikTok video intelligence. Your agents can search topics, list channel uploads, parse playlists, and pull structured transcripts with word level timestamps in seconds.

1. Install via MCP or ClawHub#

OpenClaw supports native Model Context Protocol servers as well as ClawHub skill packages. The native MCP server below is the reliable path. The ClawHub package is community published and not verified here, so prefer Option B if the skill command fails.

Option A: ClawHub skill command (unverified here)

npx clawhub@latest install transcriptdock

Option B: Native OpenClaw MCP server

Add the TranscriptDock stdio server to your OpenClaw agent configuration file (openclaw.json or config.json):

{
  "mcpServers": {
    "transcriptdock": {
      "command": "npx",
      "args": ["-y", "@transcriptdock/mcp"],
      "env": {
        "TRANSCRIPTDOCK_API_KEY": "td_live_YOUR_KEY",
        "TRANSCRIPTDOCK_BASE_URL": "https://transcriptdock.com"
      }
    }
  }
}

2. Configure environment#

Provide your TranscriptDock API key in the environment block. Replace td_live_YOUR_KEY with your actual API key from the developer dashboard.

3. Verify connection#

Start your OpenClaw agent and ask it to verify its installed tools:

Check available tools and confirm transcriptdock is connected.

Your agent should confirm tools including transcribe, search_youtube, list_channel_videos, and get_transcript.

Autonomous workflows#

Once connected, your OpenClaw agent can chain discovery and transcription tools autonomously:

• Topic research: The agent searches YouTube using search_youtube, identifies the most relevant videos, and retrieves full transcripts for synthesis.
• Channel monitoring: Using get_channel_latest_videos, your agent tracks new uploads from designated creators and summarizes them into your knowledge base.
• Tutorial to code: The agent reads transcripts from technical tutorials and converts verbal step by step instructions into executable code files.
• Playlist digest: Using list_playlist_videos, the agent walks an entire playlist or course curriculum and compiles a unified study guide.

Example agent prompts#

• "Search YouTube for recent tutorials on building autonomous agents in Python. Transcribe the top two videos and extract the core code architecture into a markdown report."
• "Get the transcript for this YouTube lecture: https://www.youtube.com/watch?v=dQw4w9WgXcQ. Summarize the main arguments and give me timestamped quotes."
• "Look up the latest five videos from the channel @veritasium. Transcribe the most recent upload and list the key questions explored in the episode."
• "Read the videos in this YouTube playlist and create an indexed syllabus with brief summaries for each session."

Troubleshooting#