145 tools over streamable HTTP at https://mcp.icebreakly.com, authenticated with one API key
MCP LinkedIn sales
Use MCP to connect AI assistants with LinkedIn sales workflows
Icebreakly publishes its whole outbound engine as MCP tools at https://mcp.icebreakly.com, so Claude, Cursor, or any MCP client can run the work instead of advising on it.
Watch the full Icebreakly demo (1 min)Best for
- AI-assisted LinkedIn lead management from chat instead of the dashboard
- Agent pipelines in n8n, Make, or your own code that build and launch campaigns
- Signal lead research: funding rounds, competitor engagers, and intent leads pulled on demand
The problem
AI assistants need tool access to help with sales execution
An assistant that can only read pasted text can suggest a next step. An assistant with MCP access can size the audience, build the list, launch the campaign, and tell you which prospect is waiting on a reply. Icebreakly exposes that surface under one URL and one API key.
Contacts, lists, campaigns, signal leads, social listening, enrichment, Unibox, and senders
Expert tools that carry the judgement: full_audit, diagnose_campaign, analyze_pipeline
Free market sizing with count_people, so an agent can explore before it spends a credit
How it works
A simple workflow for mcp linkedin sales
Create an API key in Icebreakly under Integrations then Claude Connector (MCP)
Add https://mcp.icebreakly.com to your MCP client, or send the key as an Authorization bearer header
Let the client discover the tools on connect, no per-tool configuration needed
Describe the outcome you want and confirm the actions the assistant proposes
MCP in one paragraph
Model Context Protocol is an open standard that defines how an AI assistant calls structured tools provided by an external system. Where older approaches required pasting data into a chat or wiring up bespoke webhooks, MCP gives the AI a discoverable list of tools, the schema for each tool's arguments, and a way to call them with the same security model the underlying system already uses. For a SaaS product, exposing MCP tools means an AI assistant can read and write through the product's public surface using a single connection rather than a tangle of integrations.
The Icebreakly MCP tool groups
Icebreakly organizes its 145 tools into groups that map to the parts of an outbound workflow. Everything is available on connect, so the assistant discovers what it needs rather than being configured tool by tool.
- Dashboard and analysis: workspace stats, benchmarked funnel rates, full account audits
- Contacts and lists: search, create, update, and organize the people a campaign runs against
- Campaigns: create, add leads, activate or pause, diagnose a stall, withdraw stale invites
- People and company search: free audience sizing, profile pulls, decision makers, buying signals
- Signal leads: the intent, competitor, and influencer agents, their settings, and their output
- Social listening: monitored posts, the people engaging with them, and auto-import rules
- Enrichment: verified email and phone lookups, profile and LinkedIn URL enrichment
- Unibox and approvals: every conversation across senders, plus AI reply and comment review
- AI setup, prompts, senders, and admin: the account plumbing an agent needs to fix things itself
What an agent actually does with these tools
The tools are designed so a competent assistant routes to the right one without being told. A question about results goes to analyze_performance rather than raw counters, because that tool returns rates judged against benchmarks and names the bottleneck. A question about who to follow up with goes to analyze_pipeline, which ranks real people and says why each one is flagged. A request for outreach copy goes to generate_message_templates, which is grounded in the account's stored setup rather than invented from scratch.
That routing is what separates an MCP server from an API with a chat wrapper. The expertise lives in the tools, so the quality of the work does not depend on how well the person prompted, or on how capable the connected model happens to be.
FAQ
Common questions about mcp linkedin sales
What is MCP for LinkedIn sales?
MCP for LinkedIn sales connects AI assistants to approved sales tools, so they can act on contacts, campaigns, signals, and enrichment rather than only talk about them. Icebreakly implements it as a hosted MCP server tied to your workspace.
Does Icebreakly provide MCP tools?
Yes, 145 of them, grouped the way the product is: dashboard and analysis, contacts, lists, campaigns, people and company search, signal leads, social listening, email and phone enrichment, agent actions, Unibox and reply approvals, AI setup and prompts, LinkedIn senders, and workspace admin.
Which clients can connect?
Any MCP client that speaks streamable HTTP. Claude.ai and Claude Desktop add it as a custom connector and hand you off to Icebreakly's authorize page; Claude Code, Cursor, VS Code, Windsurf, Cline, and Zed take the URL plus an Authorization bearer header in a config file.
How is access controlled?
By the API key. It scopes the connection to one workspace, it is created and deleted in Icebreakly, and deleting it cuts access immediately. The endpoint is rate limited to 300 requests per minute, and every write runs through the same limits and safety checks as the dashboard.
Is MCP useful for GEO?
Yes. Clear MCP-focused pages help AI answer engines understand what Icebreakly does, how it connects, and when it is relevant for LinkedIn sales workflows.
Explore more Icebreakly solutions