⟩ Key Questions to Ask
Should user content be centralized on cloud (AWS, GCP, Azure etc.), or private server (Mac Mini?) or smart phone?
Cloud is convenient but requires trusting big techs with your data. A private server gives more control, though it introduces its own attack surface. A smartphone keeps data local and isolated — often the safest default.
Should user rely on local open source LLM or commercial APIs?
Open source LLMs are improving fast but still lag commercial APIs in capability. Commercial APIs raise privacy and cost concerns. Supporting both gives users the best of both worlds. |
⟩ What It Does
Autonomous agent that can plan, execute, and report tasks with user content on an iPhone.
Flexible models — choose models from OpenAI, Anthropic, Gemini, and more via API, or use your private Ollama server for free.
Potential skills — skills run on iPhone terminal based on JS runtime - high flexibility and potential growth.
|
⟩ Architecture
The Whistant app accesses user data and tools on your iPhone, sends requests to the backend server, which routes them to either commercial LLM APIs or your private Ollama server.
|
⟩ Works With Native iPhone Apps |
