Canvo — AI Agent + Canvas + Linux Sandbox

Three Superpowers

Not a chatbot. A full agent with a visual canvas, a Linux environment, and tools that multiply.

The Canvas

Your agent renders. You interact.

Cards, charts, tables, forms, toggles, progress bars
Animated D3 visualizations with real-time controls
Quizzes, interactive maps, mini-games
Every response is a living workspace

Linux Sandbox

A real shell on your phone.

300+ Unix commands — grep, awk, curl, wget…
Install Python, Node.js, Git, FFmpeg, and more
Build full web apps that run locally
Persistent workspace across sessions

Unlimited Tools

If a tool doesn't exist, the agent builds it.

Web search, calendar, contacts, clipboard
File I/O, notifications, TTS, timers
Self-building tool scripts — saved and reused
Heartbeat: scheduled autonomous tasks

Always Working

Canvo isn't just reactive. Schedule autonomous agent tasks that fire throughout the day. Your agent wakes up, reads your notes, searches the web, and sends notifications — without you touching the screen.

Heartbeat: Autonomous scheduled tasks
Memory: Persistent knowledge base that grows
Proactive: Morning briefings, reminders, standups

Watch the Teaser

50 seconds of Canvo in action.

Watch on YouTube

See it in Action

Use case screenshots from the app !

Your Keys. Your Data.

No account. No subscription. Nothing touches our servers.

Bring Your Own Keys

OpenAI, Anthropic, Google Gemini, Groq, Mistral, or any OpenAI-compatible endpoint. Ollama, LM Studio, OpenRouter — all supported.

Local-First

All conversations, files, and agent memory stay on your device. Works offline with local models. No cloud sync required.

Multi-Provider

Switch between providers per session. Use the best model for the task. Google OAuth sign-in for Gemini — no API key needed.

Your AI workstation.
In your pocket.

Currently in beta. Built for developers, tinkerers, and power users.

Download Beta APK

Android 13+ · arm64-v8a · No account required