Documentation Index
Fetch the complete documentation index at: https://www.outx.ai/docs/llms.txt
Use this file to discover all available pages before exploring further.
The OutX MCP server lets AI agents interact with the full OutX API through the Model Context Protocol. Create watchlists, search posts, fetch profiles, and engage on LinkedIn, all via natural language prompts.
Install
Requires Node.js 18+ and an OutX API key. Get your key.
Setup
Claude Desktop
Add to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"outx": {
"command": "npx",
"args": ["-y", "outx-mcp-server"],
"env": {
"OUTX_API_KEY": "your-api-key"
}
}
}
}
Cursor
Add to .cursor/mcp.json in your project:
{
"mcpServers": {
"outx": {
"command": "npx",
"args": ["-y", "outx-mcp-server"],
"env": {
"OUTX_API_KEY": "your-api-key"
}
}
}
}
Claude Code
Add to .mcp.json in your project:
{
"mcpServers": {
"outx": {
"command": "npx",
"args": ["-y", "outx-mcp-server"],
"env": {
"OUTX_API_KEY": "your-api-key"
}
}
}
}
Watchlist Management
28 tools covering the full OutX API:
| Tool | Description |
|---|
create_keyword_watchlist | Monitor LinkedIn posts matching keywords |
create_keyword_watchlist_by_prompt | Create a keyword watchlist from a natural-language prompt (AI generates keywords and labels) |
list_keyword_watchlists | List all keyword watchlists |
update_keyword_watchlist | Update name, frequency, or status |
update_keyword_watchlist_prompt | Update the prompt on an existing watchlist (regenerates all keywords and labels) |
delete_keyword_watchlist | Delete a keyword watchlist |
create_people_watchlist | Track posts from specific LinkedIn profiles |
list_people_watchlists | List all people watchlists |
update_people_watchlist | Update name, frequency, or status |
delete_people_watchlist | Delete a people watchlist |
create_company_watchlist | Monitor posts from LinkedIn company pages |
list_company_watchlists | List all company watchlists |
update_company_watchlist | Update name, frequency, or status |
delete_company_watchlist | Delete a company watchlist |
Posts & Engagement
| Tool | Description |
|---|
get_posts | Search and filter posts with 15+ parameters |
get_interactions | Retrieve likes and comments for posts in a watchlist with filtering, pagination, and daily graph data |
like_post | Like a watchlist post |
comment_on_post | Comment on a watchlist post |
LinkedIn Data API
These tools wrap the async LinkedIn Data API. The MCP server handles polling automatically — you get the result directly.
| Tool | Description |
|---|
fetch_linkedin_profile | Fetch a profile by slug (name, headline, experience, education, skills) |
fetch_linkedin_posts | Fetch posts from profiles by URN |
fetch_linkedin_company | Fetch company page data by slug (industry, employee count, description) |
fetch_linkedin_company_posts | Fetch recent posts from a company page |
linkedin_send_message | Send a direct message to a 1st-degree connection |
linkedin_search_profiles | Search LinkedIn profiles by title, company, location, and more |
linkedin_fetch_connections | Fetch your 1st-degree connections with keyword search and pagination |
linkedin_like_post | Like a post by activity URN |
linkedin_comment_on_post | Comment on a post by activity URN |
get_task_status | Check status of an async task manually |
Example Prompts
Once connected, try these with your AI agent:
Social listening:
- “Create a watchlist to monitor LinkedIn posts about AI startups raising Series A”
- “Show me trending posts from my watchlists this week”
- “What are VPs posting about machine learning?”
Profile & company research:
- “Fetch the LinkedIn profile for williamhgates”
- “Get recent posts from this prospect’s profile”
- “Get company info and recent posts for OpenAI”
- “Search for VPs of Engineering at Google in the US”
- “Show me my most recently added connections”
Engagement & messaging:
- “Like the most engaging post about product management”
- “Comment ‘Great insights!’ on this post”
- “Send a message to this connection thanking them for connecting”
How It Works
The MCP server connects to the OutX API using your API key. When an AI agent calls a tool:
- The server translates the tool call into an OutX API request
- For LinkedIn Data tools, it automatically submits the async task and polls until complete
- Results are returned as structured JSON the agent can reason about
AI Agent ←→ MCP Protocol ←→ OutX MCP Server ←→ OutX API ←→ LinkedIn
Source Code
The MCP server is open source: github.com/outxai/outx-mcp-server
Partnerships
Interested in a collaboration or affiliate partnership? Reach out at support@outx.ai.
Learn More
