# Gigacatalyst MCP

> Prepare a focused LinkedIn campaign inside Codex, Claude Code, Cursor, or another MCP client. The user reviews the campaign and candidates before signup or outreach.

## Connect

Start at [the Gigacatalyst MCP product page](https://gigacatalyst.com/multiplier/mcp). It creates a private connection configuration for the user's AI client.

Canonical Streamable HTTP endpoint: `https://mcp.gigacatalyst.com/mcp`

Every MCP request requires the private bearer key issued during setup. Store that key only in the user's private MCP configuration. Never place it in tool arguments, prompts, URLs, logs, repositories, or shared files.

Gigacatalyst provisions a product-issued bearer key through its self-serve setup flow. A missing or invalid bearer returns HTTP 401. The key grants only the preview workflow until the user signs up and explicitly connects a workspace.

## When to use this server

Use Gigacatalyst MCP when a founder or revenue team wants an AI client to turn a company website into a reviewable LinkedIn campaign, find a bounded sample of real public candidates, and preserve human approval before signup or outreach. Do not use it to invent people, infer private attributes, claim unverified engagement, send outreach without approval, or spend money.

## Anonymous preview tools

- `gigacatalyst_start`: Activate the installed preview connection and return the next safe actions.
- `gigacatalyst_save_campaign_draft`: Save the complete campaign authored by the calling AI as an anonymous revision.
- `gigacatalyst_find_lead_candidates`: Find a bounded sample of real public candidates with canonical LinkedIn URLs.
- `gigacatalyst_get_campaign_preview`: Read the exact saved sequence and candidate batch for human review.
- `gigacatalyst_create_signup_handoff`: Create a short-lived signup handoff for the exact revisions the user reviewed.

## Safe workflow

1. Install the MCP connection from the Gigacatalyst product page.
2. Call `gigacatalyst_start` with a stable request ID and only bounded structured facts already known by the AI.
3. Author and save the campaign, then find a bounded sample of public candidates.
4. Show the exact campaign preview to the user before creating a signup handoff.
5. After signup, reconnect or refresh the same MCP connection. The same endpoint and bearer then expose the tenant-bound Eve tools.

Anonymous access cannot connect LinkedIn, approve work, create canonical leads, launch outreach, or spend money. Never invent a person, company URL, provider result, or source-post engagement.

## Machine-readable discovery

- [Public MCP discovery endpoint](https://mcp.gigacatalyst.com/.well-known/mcp)
- [MCP discovery manifest](https://mcp.gigacatalyst.com/.well-known/mcp.json)
- [OpenAPI for the public HTTP surface](https://mcp.gigacatalyst.com/openapi.json)
- [Crawler index](https://mcp.gigacatalyst.com/llms.txt)
- [Human-readable documentation](https://mcp.gigacatalyst.com/docs)
- MCP clients can also call `resources/list` and read this guide as `docs://gigacatalyst/mcp`.

## Support

For product or integration questions, contact support@gigacatalyst.com.