AI Coding Agent · Cloud & Mobile

Code smarter —
from any device

LEO is an AI coding agent that runs on your phone, tablet, or laptop. Compress context, orchestrate tasks, and sync via Google Drive, GitHub, or Bridge — with cloud or local models.

LEO9:41
Hey LEO, check my auth flow
Looking at your routes...
Found a gap in the token refresh handler.
Can you fix it?
curl -X POST ...
Added refresh interceptor ✅
Nice, thanks. Deploy?
Ask LEO anything...

Context Compression

LEO intelligently compacts your project context, fitting more understanding into fewer tokens. Bigger picture, smaller footprint.

Usage Monitoring

Real-time token and usage tracking across all sessions. Smart orchestration routes requests efficiently — no waste, full visibility.

Smart Orchestration

Intelligent task routing and batching. LEO decides when to compress, when to cache, and how to sequence operations for maximum efficiency.

Mobile-First

Full coding assistant on the go. Connect Google Drive or GitHub — powered by OpenCode cloud models with free tiers. Your entire project in your pocket.

Cross-Platform Sync

Use LEO on your phone, tablet, or laptop. Cloud sync via Google Drive or GitHub keeps your projects in sync everywhere, instantly.

Local via Leo Bridge

Desktop users run LEO fully local with Leo Bridge — your own local models, zero cloud dependencies, maximum privacy. Works offline.

~70%
Avg. Context Compression
Languages Supported
3
Sync Methods
100%
Usage Visibility

Choose how you connect

LEO adapts to your workflow — cloud for mobility, local for privacy.

Google Drive

Sync your projects to Google Drive and code from any browser. Uses OpenCode cloud models — the cheapest option with a free tier built in.

Cloud · OpenCode

GitHub

Connect GitHub repos and code from anywhere. Runs on OpenCode cloud models — free tier available, no local hardware needed.

Cloud · OpenCode

Leo Bridge

Desktop app that uses your local models — Ollama, LM Studio, or any OpenAI-compatible provider. Zero cloud dependencies, full privacy, works offline.

Local Desktop

Choose your experience

Start free with ads, or go ad-free for just $1.99. All core features included.

Free
Starter
$0 /mo

Full access with occasional ads. Great for trying LEO out.

  • Context compression engine
  • Usage monitoring & orchestration
  • Google Drive & GitHub sync
  • All languages & frameworks
  • Leo Bridge (desktop local)
  • Mobile access included
Get Started Free

Frequently Asked

Everything you need to know about LEO, our models, and your privacy.

OpenCode is our cloud AI provider — the most affordable option available, with a built-in free tier. When you use LEO via Google Drive or GitHub sync, all code generation and analysis runs on OpenCode's cloud models. This means you get powerful AI assistance from any device without needing local hardware. OpenCode's free tier lets you try everything before committing to a paid plan.

Cloud (Google Drive / GitHub): Best for mobile users, travelers, or anyone who doesn't want to manage local models. Runs on OpenCode's infrastructure — no downloads, no GPU needed. Your projects sync across all your devices.

Local (Leo Bridge): Best for desktop users who want full privacy, offline capability, and control over their models. Leo Bridge connects to Ollama, LM Studio, or any OpenAI-compatible provider running on your machine. Zero data ever leaves your computer.

When you use cloud sync via Google Drive or GitHub, your project files are stored in your personal cloud accounts (Google/GitHub), which are protected by their own security measures. Code sent to OpenCode's models is processed temporarily for generation and analysis.

For maximum privacy, use Leo Bridge locally — everything stays on your machine, and no data is ever sent to external servers.

Leo Bridge works with any local LLM provider. You'll need a machine capable of running your chosen models. Popular options include:

Ollama — the simplest option, works on Mac, Windows, and Linux
LM Studio — great GUI with built-in model downloads
• Any OpenAI-compatible API running locally

For smaller models like qwen3:8b or deepseek-r1:7b, even a laptop with 8GB RAM works well.

Context compression is LEO's ability to fit more project understanding into fewer tokens. Instead of sending your entire codebase to the model with every request, LEO intelligently summarizes, prioritizes, and compacts the relevant context.

This means: faster responses, lower token usage (saving you money), and the ability to work on larger projects without hitting context limits. Our average compression rate is ~70% — you get the same quality with a fraction of the tokens.

Yes! LEO supports both modes and you can switch freely. Use cloud sync when you're on the go with your phone or tablet, then pick up right where you left off on your desktop with Leo Bridge. Your projects stay in sync via Google Drive or GitHub, so nothing gets lost.

Starter (Free with ads): Full access to all LEO features — context compression, usage monitoring, Google Drive & GitHub sync, Leo Bridge, and all supported languages. Occasional ads are displayed.

Pro ($1.99/mo): Everything in Starter plus zero ads for a clean experience and priority support. If you use LEO daily, Pro is the best way to stay focused.