Connect your AI assistant to Vaam

Vaam’s MCP (Model Context Protocol) server connects AI assistants such as Claude and ChatGPT directly to your Vaam account. Your assistant can answer outreach questions, manage sequences and prospects, and help keep other systems in sync in plain language.

The server is available at:

https://app.vaam.io/api/agent/mcp

What your assistant can do

• Answer questions, such as “How many prospects are active in my sequences?”, “Which retained LinkedIn replies can I review?”, and “Am I close to my active prospect limit?”
• Manage outreach by pausing or resuming sequences, adding prospects to a sequence, fixing a prospect’s contact details, or stopping outreach to a specific person.
• Review work by listing tasks that need your attention, finding issues blocking outreach, and resolving them.
• Set up AI outreach by configuring your ideal customer profile, company info, offering, and messaging language through conversation.
• Inspect prospecting agents by reading sources, qualification results, and retained reasons for skips or holds.
• Review retained history after outreach stops, including stored replies and status changes.
• Sync data by paging through retained outreach data for other tools.

See what your assistant can do for the 34 customer tools. Read Retained replies and history before interpreting reply evidence. Collection gaps remain, even after you read every page.

Get started

1. Connect your assistant. The page includes walkthroughs for claude.ai, Claude Code, Claude Desktop, ChatGPT, and custom clients.
2. Ask a question like “What’s my outreach status?” to confirm the connection works.

Before you connect

Permissions and trust explains what a connected assistant can see. It can see your own data, your team’s data only when you ask about the team, and nothing after you disconnect it. Read it before you connect, or send it to your security team.

Building an integration?

If you are integrating programmatically rather than chatting:

• Connect a custom client. It covers login flow requirements and a scoped API key with a worked example.
• Rate limits. It explains request limits and how to handle 429 responses.
• Paging through data. It covers cursors, retained progress sync, and count calculations.

Vaam runs a second, staff-only address for its own support tools. It is separate from the address above and requires a target user for shared tools. See Permissions and trust.