Installation
Build from source in minutes. Requires CMake and a C compiler. Get started →
not quite human. A minimal, dependency-light AI assistant runtime optimized for small binaries and low memory footprint. It supports 50+ AI providers, 38 messaging channels, and 83+ tools — all in a ~1.7 MB static binary.
Installation
Build from source in minutes. Requires CMake and a C compiler. Get started →
Quick Start
Get your first chat running with the onboard wizard or a quick config. Quick start →
Providers
OpenAI, Anthropic, Gemini, OpenRouter, Ollama, and 50+ compatible APIs. See providers →
Channels
CLI, Telegram, Discord, Slack, Signal, Nostr, and more. See channels →
HuLa Programs
Structured JSON plans for multi-step tool use—policy-checked and traceable. HuLa guide →