MCP Overview
What the Icebreaker MCP server exposes, which use cases it's built for, and when to use MCP vs the REST API.
The Icebreaker MCP server makes your professional network available as a set of tools inside any AI assistant that supports the Model Context Protocol — Claude, ChatGPT, Cursor, and others.
Instead of switching between tabs, building exports, or annotating spreadsheets, you can search and act on your network using natural language, right inside the tools you're already working in.
"Find senior iOS engineers in my network in Austin"
"Who can intro me to the Head of Engineering at Vercel?"
"Search our team's network for web3 founders who've raised a seed round"What you can do
Search your network
Search across your synced LinkedIn, GitHub, and X/Twitter connections by role, skill, location, or company. Searches are scoped to your extended network (1st and 2nd degree) by default, so results are people you can actually reach.
Find warm intros
For any person in Icebreaker, the server can surface who in your network knows them and rank those connectors by trust score — giving you the strongest path to a warm introduction.
See mutual connections
Quickly surface shared connections between you and a target profile. Useful for validating a candidate, preparing for a meeting, or understanding your relationship surface before an outreach.
Search your team's combined network (Teams)
If you're on an Icebreaker Teams subscription, you can search across your entire team's collective LinkedIn, GitHub, and X/Twitter graphs in a single query. One search covers everyone's connections, not just your own.
Manage contact lists
Create, update, and query saved contact lists. Useful for tracking candidates across a search, building target account lists, or maintaining a pipeline of warm prospects.
Sync and enrich
Trigger a LinkedIn, GitHub, or X/Twitter graph sync on demand. Enrich individual people and companies by LinkedIn URL to pull contact information, job titles, and firmographic data.
MCP vs REST API
Icebreaker exposes two integration surfaces. Here's when to use each:
| MCP | REST API | |
|---|---|---|
| Best for | Conversational and agentic workflows | Programmatic integrations and custom apps |
| How you call it | Natural language via an AI assistant | HTTP requests from your code |
| Auth | OAuth (per user) | Bearer token (contact [email protected]) |
| Typical use | "Find engineers at Coinbase in my network" from Claude | Pulling profile data into your own product |
| Requires AI client | Yes | No |
Use the MCP server if you're working inside Claude, ChatGPT, or Cursor and want to interact with your network conversationally or as part of an agent workflow. Use the REST API if you're building a product or pipeline that calls Icebreaker programmatically.
Available tools
The MCP server exposes 14 tools across five categories:
Search — search_network
Warm intros and mutuals — get_intro_path_tool, get_mutuals_tool
Team network — get_team_tool, get_team_collaborators
Contact lists — list_lists_tool, get_list_tool, manage_list_tool
Sync and enrichment — queue_network_sync_tool, queue_network_sync_status_tool, enrich_person_tool, enrich_company_by_linkedin_tool, get_enrichment_result_tool, lookup_viewer_tool
For full parameter documentation and examples, see Available Tools.
Data and permissions
The MCP server only accesses data you've synced to your Icebreaker account. It cannot read connections you haven't synced, and it operates entirely within your account's visibility settings.
When you connect via OAuth, you authorize the server to act on your behalf. No data is shared with third parties. For details on what the OAuth flow requests, see OAuth.
Next steps
- Quickstart — connect the server to Claude, ChatGPT, or Cursor
- Available Tools — full tool reference
- OAuth — auth flow and scopes
Updated 3 months ago