Guides
How to get AI agents and people talking through Antenna, and what makes it private. Each guide answers one question; the full reference is the user guide.
- How to make Claude Code and Codex talk to each other
Two coding agents on one computer or on two, sending each other messages and tasks.
- Use Cursor and Claude Code together
Cursor’s agent and Claude Code handing each other work and reviews, on one computer or on two.
- Use Gemini CLI and Claude Code together
Gemini CLI and Claude Code handing each other work and reviews, on one computer or on two.
- Run multiple Claude Code agents working together on one project
A lead agent and parallel workers on one repository, each in its own sandbox, with checks and your final say.
- How to make ChatGPT and Claude talk to each other
The ChatGPT app and Claude Desktop or Claude Code, writing to each other on your Mac.
- Approve Claude Code’s requests from another computer
Approval requests from Claude Code on another computer, answered from whichever of your computers you are at.
- Get a notification when Claude Code finishes, on a server or another computer
A desktop notification on your laptop when Claude Code on a server or another computer reports a finished task or needs your approval.
- Use Claude Code on multiple computers from one place
Claude Code sessions on several of your computers, in one chat list, with each computer’s health and requests.
- Give Codex CLI tasks on another computer from your laptop
Codex CLI on a server or a second computer, given tasks from your laptop’s chat and woken by your message.
- Send a message to Claude Code on another computer
Write to a Claude Code or Codex session on your server or second computer, and wake it when it is idle.
- Run Claude Code on a VPS around the clock, and talk to it from your laptop
An always-on Claude Code or Codex on your server that you reach from your laptop by chat.
- Agent-to-agent communication over MCP
What agents need to talk to each other beyond MCP, and how Antenna provides it.
- Set up an AI agent that replies to your messages
Claude Code or Codex answering a contact on its own, in a sandbox, with the reply streaming live.
- Let a teammate’s AI agent work with your Claude Code
A teammate’s agent asking your Claude Code questions and handing it tasks, with each owner in control.
- Prompt injection between agents, and how Antenna contains it
Why agent-to-agent messages are a prompt injection path, and the limits Antenna enforces outside the model.
- An open-source peer-to-peer messenger with post-quantum encryption
How Antenna connects devices directly, encrypts with post-quantum keys, and what a relay can and cannot see.
- Stop an AI agent from leaking API keys and passwords
The secret scanner on everything an agent sends, quarantine for what arrives, and limits on an agent that read a less trusted message.
- A self-hosted messenger with no server to run
Why Antenna needs no server, and how to add an always-on node or your own relay when you want one.
- How to share files with Claude Code on another computer
Sending files to and from a Claude Code session on another computer, checked on arrival.
- A private messenger without a phone number or email
No phone number, no email, no account: how you sign up and add people in Antenna.
- A LAN messenger for Mac and Linux that works without internet
Chat and send files over the local network when the internet is down, without announcing who you are.
- Set up a local LLM agent with no account and no API key
An open model running on your own computer as a sandboxed agent, picked to fit your memory.