Last verified
Orchestrate Claude Code, Codex, Grok, and Gemini in one app by reusing the subscriptions you already pay for, on desktop and mobile.
If you pay for Claude Pro or Codex and want to switch between them for different tasks without leaving the same interface, Happy unifies session management without adding a new subscription layer. You bring your credentials; Happy manages the sessions.
You use only one AI coding provider and have no need for multi-model switching. A single-provider workflow is simpler without the extra coordination layer.
Deploy the team relay server and invite colleagues into the same running agent session. Each participant authenticates as themselves and can steer the agent, which makes code review and pair programming with AI agents possible in real time.
You work solo or your team collaboration is already covered by a shared workspace tool. The multiplayer server requires infrastructure setup that solo developers do not need.
Start a Claude Code or Codex session at your desk and pick it up on iOS or Android without losing context. The mobile apps are end-to-end encrypted and rated 4.9 and 5.0 stars respectively, with the same session state you left on the desktop.
You work exclusively at a desk and do not need mobile access. The mobile sync layer adds relay server dependency with no benefit for desktop-only workflows.
With MIT licensing across all four repositories (mobile app, desktop app, agent runtime, relay server), you can fork any component and ship your own build. The agent runtime (happy-agent) is the open source runtime the desktop app uses, making it a starting point for custom agent integrations.
You need a stable, fully polished product and do not have engineering bandwidth to work with an actively developed codebase that has over 1,000 open issues.
AI coding tools have splintered into a subscription maze. Claude Code, Codex, Grok, and Gemini each come with their own interfaces, their own session contexts, and their own apps. Developers building seriously with AI spend time switching between these environments, losing agent context at each boundary, and still paying for the underlying model subscriptions regardless.
Existing interface layers add a third bill on top: tools like Cursor and GitHub Copilot charge a per-seat fee for their own model wrappers. If you are already subscribed to Claude Pro, Codex, and Grok, you are paying for model access that Cursor's interface reroutes through its own infrastructure anyway. The waste is the interface layer itself, duplicating subscription costs you already carry.
Run Claude Code, Codex, Grok, and Gemini from the same app. Assign each task the model that fits best, and delegate to subagents on any of them. Gemini runs through Antigravity; OpenCode and any Agent Client Protocol agent can plug in as an additional provider.
You authenticate directly with your Claude, Codex, or Grok accounts. Happy acts as the coordination layer; it does not resell model access or add a per-seat fee on top of your subscriptions. Gemini connects through your Google account via Antigravity.
The iOS and Android apps connect to the same agent sessions you started on the desktop. Data moving between your devices is encrypted in transit. The iOS app has 4.9 stars from over 1,000 App Store ratings; the Android app has 5.0 stars from over 3,100 Google Play reviews.
Deploy a team relay server and invite colleagues into the same agent session. Each participant signs in as themselves, shares the same project context, and can steer the running agent. Built for teams that want shared AI coding sessions without screen sharing.
Install the happy npm package to wrap Claude Code or Codex directly in your terminal. Start an agent with 'happy claude' for Claude Code or 'happy codex' for Codex, then continue the same session from your phone through the relay server. No desktop app required.
Both tools make AI coding agents productive on your machine, but they work differently. Cursor is a complete code editor that bundles model access and charges a per-seat monthly subscription. Happy is a coordination layer only, routing your sessions through the Claude, Codex, or Grok subscriptions you already hold, with no per-seat fee for the orchestration layer.
| Feature | Happy | Cursor |
|---|---|---|
| License | MIT | Proprietary |
| Self-hosting | Relay server (open source) | No |
| Model access | Your existing accounts | Bundled per-seat subscription |
| Mobile app | iOS and Android (encrypted) | No |
| Multiplayer sessions | Yes, team server | No |
| Platform fee | None | Monthly per-seat |
Happy is the better choice when you already hold active Claude Pro or Codex subscriptions and do not want to pay a second interface fee on top of them. It also wins for teams needing multiplayer sessions or mobile continuity. Cursor is worth considering when you want a tightly integrated editor environment with model access included, or when you do not hold individual model subscriptions and prefer an all-in-one coding environment.
GitHub Copilot provides inline completions and chat inside VS Code, JetBrains, and other editors at a per-seat monthly fee. Happy manages full agent sessions where Claude Code or Codex run end-to-end coding tasks autonomously, with multi-model switching and mobile continuity.
| Feature | Happy | GitHub Copilot |
|---|---|---|
| License | MIT | Proprietary |
| Self-hosting | Relay server | No |
| Model | Your Claude/Codex/Grok accounts | GitHub's hosted models |
| Mobile access | iOS and Android apps | No dedicated app |
| Agent sessions | Full multi-step autonomous agents | Chat completions and inline |
| Platform fee | None | Monthly per-seat |
Happy fits developers who use AI for full autonomous coding tasks, not just completions and chat. GitHub Copilot is the simpler choice if your primary use is autocomplete and editor chat inside VS Code without managing separate model subscriptions. If you already pay for Claude Pro and want to run long autonomous sessions across desktop and phone, Happy serves a different workflow than Copilot does.
Install the desktop app on macOS via Homebrew, or set up the CLI for terminal-first workflows.
```bash
brew install --cask slopus/tap/happy
npm install -g happy
```Yes, Happy itself is MIT licensed and free. Your costs are the model subscriptions you use (Claude, Codex, Grok, or Gemini via Antigravity); Happy adds no per-seat fee for the coordination layer. The desktop app is available for macOS, Windows, and Linux; the mobile app is on iOS and Android.
Yes. The relay server (happy-server) and agent runtime (happy-agent) are open source and self-hostable. Self-hosting the relay keeps your session sync data on your own infrastructure. The desktop and mobile apps connect to a self-hosted relay by configuration. Full self-hosting requires running multiple components: the relay server and optionally the agent runtime alongside your model connections.
Happy supports Claude Code (Anthropic), Codex (OpenAI), Grok (xAI), and Gemini via Antigravity. Any Agent Client Protocol (ACP) agent can also plug in as a provider. Each model requires its own active subscription from the respective provider; Happy acts as the coordination layer on top of those accounts.
Cursor and GitHub Copilot bundle their own model access and charge a per-seat monthly fee for their interface layer. Happy acts as a coordination layer, routing your sessions to the model accounts you already hold without adding model charges. The tradeoff is that Happy requires individual model subscriptions for each provider you want to use.
Happy has a desktop app for macOS (Intel and Apple Silicon), Windows, and Linux, available via Homebrew on macOS or direct download. The mobile app is on iOS (App Store, 4.9 stars) and Android (Google Play, 5.0 stars). An npm CLI (npm install -g happy) is available for terminal-first workflows that wrap Claude Code and Codex directly.
Alternative to
Grok Bot
Self-hosted AI agent: persistent memory, multi-channel chat, model choice.
Alternative to
Browserbase
Records portal tasks once; replays with no model and no bot detection
Alternative to
Conductor
Track coding agent sessions with cost, plan limits, and PR pipeline
Alternative to
Cursor
Local-first AI agent desktop app for Claude Code, Codex, and Pi.
Alternative to
Braintrust
Automated LLM evals, red teaming, and AI vulnerability scanning.
Alternative to
Together AI
High-throughput LLM inference server with OpenAI-compatible API
Sponsors
Alternative to
Superhuman
AI email assistant that drafts replies and clears your inbox
Alternative toClaude Managed Agents
Open-source agent deployment API. Self-host or use Omnara Cloud.
Alternative toGrok Bot
AI teammates you own: your keys, your model, your machine.
The projects we add and every new guide. Free, unsubscribe any time.
Claude Code to 
OpenCode
Jira to 
Windshift
Notion to 
AppFlowy
Slack to 
Mattermost
Linear to 
GitLab
Wispr Flow to 
Handy
Arc to 
Zen Browser
CapCut to 
Concat
Grok Bot to 
rakazo
Superhuman to 
Inbox Zero
Airtable to 
AITable
Muse to 
OpenClaw
Railway to 
Coolify
GitHub Copilot to 
Cline
Datadog to 
Netdata
Microsoft Teams to 
Rocket.ChatConfluence to 
Docmost
Pinecone to 
Qdrant
Zapier to 
n8n
Devin AI to 
OpenHands
Firebase to 
Supabase
1Password to 
KeePassXC
Contentful to 
Strapi
Adobe Premiere Pro to 
Shotcut
