OpenClaw YouTube Comments Plugin

OpenClaw plugin for managing YouTube channel comments — scan for new comments, generate AI replies, and post with an approval workflow via Telegram.

Features

• Interactive mode — review and approve each reply one by one via Telegram
• Dry-run mode — preview all generated replies without posting
• Auto mode — automatically post replies to all new comments
• Multiple identities — switch between personas (e.g. volkova, openprophet)
• Thread support — continue conversations in existing comment threads
• Dual AI backend — use OpenClaw's connected model or direct Gemini API
• Background polling — automatic periodic checks for new comments
• /yt slash command — quick status without consuming AI tokens

Installation

1. Clone this repository into your OpenClaw plugins directory
2. Install dependencies:

   npm install

3. Build:

   npm run build

4. Add to your openclaw.json:

   {
     "plugins": {
       "youtube-comments": {
         "path": "/path/to/openclaw-youtube",
         "config": {
           "channelId": "UCxxxxxxxxxxxxxxxx",
           "oauthCredentialsPath": "/path/to/client_secret.json"
         }
       }
     }
   }

Configuration

Key	Type	Default	Description
channelId	string	required	YouTube channel ID to monitor
videoIds	string[]	—	Specific video IDs (overrides maxRecentVideos)
maxRecentVideos	number	5	Recent videos to check
maxCommentsPerVideo	number	50	Max comments per video
maxCommentAgeDays	number	7	Ignore older comments
minCommentLength	number	3	Skip short comments
defaultIdentity	string	"volkova"	Default persona
replyDelayMin / replyDelayMax	number	10 / 60	Auto-mode delay (seconds)
oauthCredentialsPath	string	—	Path to Google OAuth client_secret.json
oauthTokenPath	string	auto	Where to store OAuth token
geminiApiKey	string	—	Direct Gemini API key (optional)
geminiModel	string	"gemini-2.0-flash"	Gemini model when using direct API
pollIntervalMinutes	number	120	Background polling interval (0 = disabled)

OAuth Setup

1. Create a project in Google Cloud Console
2. Enable the YouTube Data API v3
3. Create OAuth 2.0 credentials (Desktop app type)
4. Download client_secret.json
5. Set oauthCredentialsPath in plugin config
6. On first use, a browser window will open for authentication

Usage

Via Telegram (or any OpenClaw chat interface):

• "Check comments" — interactive scan with approval for each reply
• "Dry run" — preview replies without posting
• "Auto reply all" — post all replies automatically
• "Check comments as openprophet" — use a different identity
• /yt — quick status
• /yt scan — count new comments
• /yt identities — list available identities

Identities

Identity files in identities/ define the persona and tone:

• volkova.txt — Nastya / VOLKOVA channel persona
• openprophet.txt — Eugene / tech-oriented persona

Format: metadata header (name, handle, channel) separated by --- from the prompt text.

Development

npm run dev      # Watch mode
npm run test     # Run tests
npm run lint     # Type check

Architecture

src/
  index.ts              # Plugin entry — registers tools, commands, service
  auth.ts               # OAuth 2.0 flow for YouTube
  youtube.ts            # YouTube Data API wrapper
  state.ts              # Replied comments tracking (JSON)
  identities.ts         # Identity loader + prompt templates
  reply-generator.ts    # Dual backend: OpenClaw model / Gemini API
  types.ts              # TypeScript interfaces

License

MIT