HeyLead — Claude Code MCP server for LinkedIn outreach
Finds people on LinkedIn, writes to them in your style, follows up and sorts replies. Works inside Claude, Cursor or any MCP client.
claude mcp add --transport http heylead https://heylead.dev/mcpWhat HeyLead — AI agent for LinkedIn outreach does
HeyLead hands your agent a set of LinkedIn tools: build a buyer profile, locate matching people, send invitations, chase the acceptances and read incoming answers. A campaign starts life as a draft and only goes live when you launch it. Sending happens from your own LinkedIn account at a human pace — no more than 20 invitations a day and 100 a week on a free account, Monday to Friday from 08:00 to 22:00 your time. When LinkedIn pushes back, the server slows down and picks up again later.
On the hosted setup, every opening DM and follow-up waits for a person to approve it by default, and the approved text is what goes out. Invitations, InMail and replies are never held. Autopilot is one switch away in sending settings.
Self-hosting is an option too. You bring a Unipile account, which is what actually talks to LinkedIn, plus your own LLM API key — model calls are billed to you. Set the server address and the Unipile key with heylead config set unipile_api_url <DSN> and heylead secrets set unipile_api_key; the key lands in the OS keychain rather than the config file. Local runs go through uvx, so uv has to be installed.
What it helps with
- locate CTOs at fintech startups for a target ICP
- build a buyer persona with pain points and objections
- send LinkedIn invitations without blowing your account limits
- follow up with people who accepted but never answered
- sort incoming replies and surface hot leads
Example prompts for the agent
- Send outreach to this campaign
- Check my LinkedIn replies
How to install HeyLead — AI agent for LinkedIn outreach
1Platform CLI: remote server
claude mcp add --transport http heylead https://heylead.dev/mcp2Platform CLI
claude mcp add heylead-ai-agent-for-linkedin-outreach -- uvx heyleaduvx runs the command named after the package; if the README names another one, use uvx --from heylead <command>
1Platform CLI: remote server
codex mcp add heylead-ai-agent-for-linkedin-outreach --url https://heylead.dev/mcp2Platform CLI
codex mcp add heylead-ai-agent-for-linkedin-outreach -- uvx heyleaduvx runs the command named after the package; if the README names another one, use uvx --from heylead <command>
Config file
{
"mcpServers": {
"heylead-ai-agent-for-linkedin-outreach": {
"url": "https://heylead.dev/mcp"
}
}
}{
"mcpServers": {
"heylead-ai-agent-for-linkedin-outreach": {
"command": "uvx",
"args": [
"heylead"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from heylead <command>
Config file
{
"servers": {
"heylead-ai-agent-for-linkedin-outreach": {
"type": "http",
"url": "https://heylead.dev/mcp"
}
}
}{
"servers": {
"heylead-ai-agent-for-linkedin-outreach": {
"type": "stdio",
"command": "uvx",
"args": [
"heylead"
]
}
}
}uvx runs the command named after the package; if the README names another one, use uvx --from heylead <command>
1Platform CLI: remote server
gemini mcp add -t http heylead-ai-agent-for-linkedin-outreach https://heylead.dev/mcp2Platform CLI
gemini mcp add heylead-ai-agent-for-linkedin-outreach uvx heyleaduvx runs the command named after the package; if the README names another one, use uvx --from heylead <command>
Step by step, with commands for each platform: How to add an MCP server to Claude Code, Codex, Cursor, VS Code, Gemini CLI and more
Questions about HeyLead — AI agent for LinkedIn outreach
How do I install HeyLead in Claude Code?
Run claude mcp add --transport http heylead https://heylead.dev/mcp, then type /mcp, pick heylead and choose Authenticate to sign in.
What does HeyLead do?
It runs LinkedIn outreach for you: searches for people matching your ICP, writes messages in your voice, sends invitations and follow-ups, and triages replies into hot leads.
What does self-hosting require?
A Unipile account with its DSN and API key, plus an LLM API key that you pay for. Local runs use uvx heylead, which needs uv installed.
