AI applications

ApuOS provides an AI workspace for local and cloud model services. You choose the data path and credentials for each connection.

* Conversation — multi-turn text, image and audio conversations where the selected model supports them. * Local models — managed GGUF and llama.cpp runtimes when the device has enough CPU, memory or accelerator resources. * Speech and OCR — offline speech recognition, speech synthesis and image text recognition after the required models are installed. * Knowledge retrieval — keyword and local semantic search over indexed notes; chat does not automatically read every file. * AI gateway — model aliases, routing, application keys, quotas and request records. * Agents and automation — registered systemd/Docker services, skills, MCP connections and scheduled workflows.

Remote model services may receive prompts and attachments. Review the provider policy and configure only the credentials and data you intend to share.