Mobile sandboxes for your AI agents controlled over REST API and MCP
Stage
Not given
Not given
Founded
Not given
Not given
Based in
US
US
Pricing model
Open source
Paid
Pricing
Not given
Cloud Android 14: $0.01 per phone-minute; iPhone and iPad: $0.05 per phone-minute; Real Android devices: $0.03 per phone-minute; Dedicated devices: from $79 per month. Billed per second with one-minute minimum. Boot time and parked phones free. New accounts start with $2 credit.
Key features
Harpe: chooses what an AI agent reads under a token budget with consistent results
Kibisis: keeps agent memory on your hardware with AES-256-GCM encryption
Argus: append-only record that detects edits with cryptographic verification
No network code by default for local and disconnected use
MIT License with published code, benchmarks and raw results
Cloud Android 14 phones (arm64)
iPhone and iPad sandboxes
Real Android devices
REST API and MCP server
Screen as numbered text elements
Batch actions (tap, type, swipe up to 20 per call)
Park and resume phones with state preserved
Live view with step-by-step recording
App installation and management
Screenshot and text observation modes
What makes it different
Measured empirical validation across 6,111 questions outperforming full-context models
Zero plaintext on disk with encrypted vaults matching plaintext accuracy
78.3% fewer tokens sent to the model with a 1,200-token budget per turn
Preserves conversational chronology and relational qualifiers under token budgets
Same API and MCP for Android, iPhone, iPad and real devices
Agents interact with real apps like people do
Boot time and parked phones are free
Works with Claude Code, Codex, Cursor and any HTTP MCP client