happier-dev/happier

Happier – Multi‑device control room for AI coding agents

What it is – Happier is an open‑source, self‑hostable platform that lets you run, coordinate and steer many LLM‑based coding assistants (Claude Code, Codex, OpenCode, Gemini, Pi, etc.) across all of your computers, servers and mobile devices. It provides a unified UI (desktop app, web, iOS/Android) plus a CLI/daemon that talks directly to the model providers.

Key features

• Parallel sessions – launch any number of independent coding‑agent sessions on any machine you own. Each session runs in its own Git worktree, so you get a real checkout/branch per task.
• Sub‑agents & hand‑off – a session can spawn a secondary agent (e.g., Claude asks Codex for a review) and you can switch the primary model mid‑session without losing context.
• Cross‑device sync – the same session appears on desktop, phone and browser. You can type in a terminal, approve changes on your phone, and view diffs in the app.
• Built‑in Git workflow – stage, commit, push, create PRs, and browse line‑by‑line diffs directly from the UI.
• Voice assistant – optional voice input via ElevenLabs or a local pipeline.
• End‑to‑end encryption – all session content is encrypted on your devices before reaching the relay server; the server only sees routing metadata.
• Self‑hosting – a single command (happier relay host install) runs a relay service you control, with Docker/Proxmox options.
• Extensible via MCP – actions are defined once and can be invoked from the UI, slash commands, voice, or any external MCP host.

How it works

1. Install – desktop app (sets up CLI + daemon) or just the CLI (curl … | bash or npm i -g @happier-dev/cli).
2. Add devices – sign in on phone/tablet; sessions, machines and settings sync automatically.
3. Start a session – happier (Claude Code) or happier codex etc., picking a machine and folder. The daemon launches the chosen agent on that machine and the relay syncs the session.
4. Interact – send prompts, queue or edit messages, review diffs, commit code, or hand the session to another machine.

Who it’s for – developers who already have access to LLM coding models and want a coordinated workflow across many machines and devices, especially teams handling many parallel code‑generation tasks.

License & community – MIT‑licensed, built in Switzerland. Open source on GitHub, with Discord, discussions and contribution guidelines.