Moltis

Moltis is a self-hosted personal AI assistant built in Rust that runs as a single binary with no runtime dependencies. It supports local LLMs with automatic download and setup, multiple cloud providers, and offers multi-channel access through Web UI, Telegram bot, and JSON-RPC API. The assistant features sandboxed execution environments for safer automation and long-term memory with hybrid vector and full-text search.

• Single Binary Deployment - Download and run one self-contained binary with no runtime dependencies, available via Homebrew, Cargo, Docker, or direct download for Debian, Fedora, Arch Linux, Snap, and AppImage.

• Local LLM Support - Run your own models locally with automatic download and setup included, supporting offline operation with GGUF embeddings and provider fallback chains.

• Multi-Channel Access - Connect through Web UI, Telegram bot, JSON-RPC API, or Mobile PWA with push notifications and multi-device sync capabilities.

• Sandboxed Execution - Run browser sessions and shell commands in isolated Docker containers or Apple Container isolation for safer automation with SSRF protection.

• Long-term Memory - Hybrid vector and full-text search enables your agent to remember context, with file watching, live sync, and session export features.

• MCP Server Support - Extend functionality with MCP tool servers, skills, and hooks supporting Stdio or HTTP/SSE with auto-restart capabilities.

• Voice Integration - Talk to your assistant with multiple cloud and local TTS/STT providers for speech-to-text and text-to-speech functionality.

• Security Features - Includes passkeys (WebAuthn), scoped API keys, secrets zeroed on drop, human-in-the-loop approval, origin validation, and no unsafe code.

• Observability - Built-in Prometheus metrics, OpenTelemetry tracing, structured logging, per-provider charts, and real-time WebSocket monitoring.

To get started, install Moltis using the quickstart script, Homebrew, Cargo, or Docker. Open the web interface at localhost:13131 to complete setup with passkey or password authentication. Configure your LLM providers and channels, then start chatting with voice, map, and tool use capabilities.