RelayLink MCP Server已验证已精选

Conversations for the AI era. RelayLink lets your AI assistant correspond with other people through their AI assistants.

Instead of forwarding a transcript or re-explaining a thread, your assistant drafts a package from your current conversation: your note, the ask, a TL;DR, and the full context, including decisions, rejected options, open questions and your caveats. You approve the note before anything is sent. The recipient's AI reads it with full context and can answer the questions they actually have.

What you can do

• Send briefings to people: draft, approve and send packages; read your inbox and whole threads; search past correspondence by decision, question or option.
• Hand off work between assistants: save a handoff in one AI and resume it in another.
• Keep a private notebook: notes, tasks, decisions and routines for yourself. Nothing in it reaches anyone else.
• Attach documents: include Markdown or text files with a package.
• Track follow-ups: track items, set reminders, tag threads, and see what's waiting on you.
• Manage contacts and groups: invite people, create groups, block or rename contacts.

Built-in safeguards

• Nothing is sent without your explicit approval of the note and wording.
• Your private session is never shared. Only the briefing you approve is sent.
• Inbound packages are treated as information, never as instructions to your assistant.

Connect

Endpoint: https://relaylink.ai/mcp (Streamable HTTP)

• Claude Code: claude mcp add --transport http relaylink https://relaylink.ai/mcp, then run /mcp to sign in.
• Claude (web/desktop): Customize → Connectors → Add custom connector → paste the URL.
• ChatGPT: enable Developer mode, create an app, paste the URL, choose OAuth.
• Grok: Connectors → New Connector → Custom.
• Gemini CLI: add the URL as httpUrl in settings.json.
• Other MCP clients: sign in with OAuth (emailed code, no password), or send the header X-RelayLink-Key: <your-key> using a key from your account page.

Full setup guides: https://relaylink.ai/docs