Ollama and LM Studio on your iPhone

Yes: PocketHam. Ollama keeps running on your Mac, Windows PC or Linux box, and the PocketHam iPhone app chats with it from anywhere: home Wi-Fi, cell, a hotel, the office. The same app works with LM Studio, llama.cpp, MLX-LM, vLLM, Jan, LocalAI and any local server that speaks the OpenAI-style API. Chatting with your own models is free.

1. Install the free PocketHam desktop app on the computer that runs Ollama (download).
2. Open it. It finds Ollama and every model you've pulled on its own, on any port.
3. Install PocketHam on your iPhone and scan the pairing code on your computer's screen.
4. Pick a model and start chatting. Leave home and keep going: it connects directly when it can and through a sealed relay when it can't.

None of them. PocketHam talks to Ollama on the computer itself, so Ollama stays on its default local-only address (no OLLAMA_HOST=0.0.0.0, nothing on port 11434 open to the internet). Your phone connects to PocketHam end to end encrypted, through home routers and firewalls as they are. Only PocketHam's own traffic goes home; the rest of your phone works as normal.

• Chat with any model on any computer you pair, and switch models mid-conversation
• Keep your chat history on your own computer
• See what each AI app is using, start or stop it and free up memory with a tap
• Dictate and hear replies in natural voices made by your own computer (Pro)

No cloud AI is involved. Your prompts and replies travel end to end encrypted between your phone and your computer, with keys only your devices hold. No account, no analytics. The details are on the privacy page.