Skip to main content

​
Multi-agent communication for Claude, Gemini, Codex, and OpenCode

Agents running in separate terminals are isolated from each other. Context doesn’t transfer, decisions get repeated, file edits collide. hcom connects your existing AI tools. When one agent edits a file, runs a command, or sends a message, other agents find out immediately. hcom demo

​
Why hcom?

Real-time messaging

Agents message each other with @mentions, intents, threads, and broadcast. Messages arrive mid-turn or wake idle agents immediately.

Collision detection

If 2 agents edit the same file within 30 seconds, both get collision notifications automatically.

Event tracking

Query event history for file edits, commands, status changes, and lifecycle events. Subscribe to react to activity.

Cross-terminal spawning

Spawn, fork, resume, or kill agents in new terminal panes. Works with kitty, wezterm, tmux, and more.

Transcript sharing

Read another agent’s conversation history with configurable detail levels and range selection.

Cross-device sync

Connect agents across machines via MQTT relay for distributed workflows.

​
How it works

Messages arrive mid-turn or wake idle agents immediately. Hooks record activity and deliver messages: 
agent → hooks → db → hooks → other agent

​
What you can do

Tell any agent:
“send a message to claude”
“when codex goes idle send it the next task”
“watch gemini’s file edits, review each and send feedback if any bugs”
“fork yourself to investigate the bug and report back”
“find which agent worked on terminal_id code, resume them and ask why it sucks”

​
Agent capabilities

Agents can:
  • Message each other with intents, threads, broadcast, and @mentions (hcom send)
  • Read transcripts from other agents with ranges and detail levels (hcom transcript)
  • View terminal screens and inject text/enter for approvals (hcom term)
  • Query event history for file edits, commands, status, and lifecycle (hcom events)
  • Subscribe and react to each other’s activity (hcom events sub)
  • Spawn, fork, resume agents in new terminal panes (hcom N claude|gemini|codex|opencode, hcom r, hcom f)
  • Kill agents and close their terminal panes (hcom kill)
  • Build context bundles with files, transcript, and events for handoffs (hcom bundle)

​
Multi-agent workflows

Included workflow scripts:
An agent (or background clone) writes an honesty self-eval. A spawned calibrator reads the target’s transcript independently. A judge compares both reports and sends back a verdict via hcom message.
A judge spawns and sets up a debate with existing agents. It coordinates rounds in a shared thread where all agents see each other’s arguments, with shared context of workspace files and transcripts.
Headless agent reads every file in a path, subscribes to file edit events to stay current, and answers other agents on demand.
Create your own workflows by prompting: “read hcom run docs then make a script that does X”

​
Supported AI tools

Claude Code, Gemini CLI, Codex, OpenCode — messages are delivered automatically, events are tracked. Any AI tool that can run shell commands — send and receive messages manually with hcom start. Any process — fire and forget: hcom send <message> --from botname

​
Terminal support

Spawning works with any terminal emulator. Run hcom config terminal --info to set up a custom one. Closing/killing works with kitty, wezterm (auto open/close panes), and tmux (background). 
hcom config terminal kitty    # set terminal preference

​
Next steps

Installation

Install hcom with curl, pip, or uv

Quickstart

Launch your first agents and send messages

Build docs developers (and LLMs) love

Get started for freeTalk to us