IonClaw

IonClaw is a C++ AI agent orchestrator for running AI agents natively and offline. It is built for servers, desktop systems, Raspberry Pi, and mobile devices, and its core claim is that the same codebase compiles to native builds across those targets. The product presents itself as suitable for keeping AI work on-device, with the phone use case described as turning a smartphone into a personal AI assistant.

It supports local GGUF models through a built-in llama.cpp engine. The feature set also includes multi-agent orchestration, a browser-based control panel, a skills extension system, multi-provider LLM support, and full MCP client and server integration inside a single native build. The site says the software runs without any runtime, container, or external dependency, and that the compiled binary itself has zero runtime dependencies.

IonClaw is delivered as a native build rather than a hosted service. The supported targets named on the site are Linux, macOS, Windows, iOS, Android, Raspberry Pi, watchOS, tvOS, and cloud servers. The quick-start flow shown on the site involves cloning the repository, building, installing, initializing a project, and starting a server that opens a web panel at localhost. The project is fully open-source and MIT-licensed.

Pricing is stated plainly: IonClaw is free and open source, with no license to buy, no subscriptions, no seat limits, and no feature paywalls. The only possible cost mentioned is usage from a cloud LLM provider chosen by the user; local models are described as free and fully offline.