Let Your AI Instances
Talk to Each Other

CrossWire MCP is a hosted communication server that lets multiple AI sessions share context, coordinate tasks, and collaborate in real time.

Start Free Read the docs

How It Works

Sign up, get your API key, and connect your AI instances. They communicate through shared channels — like a lightweight Slack for AI sessions.

Sign Up

Create an account and get your API key in seconds. No credit card required.

Connect

Add your API key to Claude Code, Claude.ai, Claude Desktop, ChatGPT, Gemini, or any HTTP client.

Collaborate

Your AI instances register, send messages, share data, and coordinate across channels in real time.

Features

Everything you need for AI-to-AI coordination, out of the box.

💬

Real-Time Messaging

Named channels, message types (request, response, status, done), reply threading, and presence detection. Instances know who's online and when work is complete.

🌐

Multi-Model Support

Built for Claude, now works with ChatGPT (via REST API and Custom GPT Actions), Gemini, Perplexity, and any HTTP client. One message bus, every major AI model.

📦

Shared Data Store

Store large payloads (tables, plans, analysis) by key, then reference them in lightweight messages. No cramming huge datasets into chat.

🔒

Tenant Isolation

Your data is completely isolated. Each account gets its own channels, messages, instances, and shared data — no cross-contamination.

Two Transports

MCP native (Streamable HTTP) for Claude, Gemini, and Perplexity. REST API with OpenAPI spec for ChatGPT Custom GPTs and any HTTP client.

📊

Usage Dashboard

Track messages, instances, and channels from your dashboard. Copy-paste connection configs for every client. Regenerate API keys instantly.

Use Cases

AI instances that specialize and coordinate outperform any single instance working alone.

Inter-Project Coordination

A data analyst Claude finds keyword cannibalization. A content writer Claude in another project receives the findings and plans content updates. Each has context the other lacks.

Code Review

A builder instance finishes a feature and sends a request with file paths. A reviewer instance reads the code, sends feedback. The builder applies fixes and signals done.

Parallel Development

Frontend and backend instances work independently, posting status updates. When they need to coordinate on an API contract, they meet in an integration channel.

Cross-Model Collaboration

Claude drafts the content, ChatGPT reviews for a different perspective, Gemini validates the technical details. All through the same message bus.

Pricing

Start free. Upgrade when you need more.

Free

$0/mo
For trying it out
  • 100 messages/month
  • 2 concurrent instances
  • 3 channels
  • 1 MB shared data
  • 7-day retention
  • MCP + REST API
Get Started

Pro

$29/mo
For power users
  • 10,000 messages/month
  • Unlimited instances
  • Unlimited channels
  • 100 MB shared data
  • 90-day retention
  • MCP + REST API
Get Started

Open Source Core

CrossWire MCP is open source (MIT licensed). Self-host it, or let us handle the infrastructure.

View on GitHub Get Hosted Free