Social listening plugin for Octolens. Monitor brand mentions, competitor activity, and industry conversations across 13+ platforms. Query mentions, run analytics, manage keywords and feeds via REST API tools, or connect the Octolens MCP server for natural-language workflows.
Social listening for the Vellum assistant. Monitor brand mentions, competitor activity, and industry conversations across 13+ platforms using Octolens.
Tracks mentions of your brand, competitors, and industry terms across Reddit, X, GitHub, Hacker News, LinkedIn, YouTube, Bluesky, TikTok, podcasts, newsletters, and more. Each mention includes AI-assigned sentiment, relevance score, and topical tags.
Four tools for programmatic access, scheduled jobs, and structured queries:
| Tool | Purpose |
|---|---|
octolens_mentions | List, get, and update mentions with filters |
octolens_analytics | Sentiment distribution, volume over time, per-keyword and source breakdowns |
octolens_keywords | Create, update, pause, delete keywords; list tags and suggestions |
octolens_feeds | Create and manage saved filter views and notification destinations |
Setup: Create an API key in Octolens (Settings > API), then tell your assistant "set my Octolens API key to ". The key is stored locally in the plugin data directory.
For natural-language workflows, connect the Octolens MCP server directly:
assistant mcp add octolens -t streamable-http -u https://app.octolens.com/api/mcp/v2
assistant mcp auth octolens
Uses OAuth (no API key needed). Opens a browser tab on first use for sign-in.
assistant plugins install octolens
Pick one or both methods:
REST API (for tools and scheduled workflows):
MCP server (for interactive natural-language queries):
Tell your assistant: "connect my Octolens MCP server"
Or run it yourself:
assistant mcp add octolens -t streamable-http -u https://app.octolens.com/api/mcp/v2
assistant mcp auth octolens
A browser tab opens for OAuth sign-in. No API key needed.
Once connected, try these:
Ask your assistant to create a scheduled job that calls octolens_mentions daily:
This uses the REST API tools (not MCP) since scheduled jobs run headlessly.
octolens/
├── package.json
├── config.json
├── hooks/
│ └── init.ts — setup checks and guidance on load
├── tools/
│ ├── octolens_mentions.ts — list, get, update mentions
│ ├── octolens_analytics.ts — sentiment, volume, per-keyword, source
│ ├── octolens_keywords.ts — CRUD keywords, tags, suggestions
│ └── octolens_feeds.ts — CRUD feeds, destinations, Slack search
├── skills/
│ └── social-listening/
│ └── SKILL.md — workflow guide for both paths
└── src/
└── api.ts — shared REST client (auth, rate limits, errors)
https://app.octolens.com/api/v2{ error: { code, message, status, details } }Full docs: octolens.com/docs
MIT