MyLLM — Private AI Assistant

Opticell Ltd.

In-app purchases

About this app

Your AI. Your server. Your data.

MyLLM is a private AI assistant that puts you in control. Connect it to your own Ollama server at home, bring your own API key, or run open-source models entirely on your phone — no account, no tracking, no middleman reading your chats.

Private by design • Chats go only to the server YOU configure — or never leave the device at all • No sign-up, no analytics, no data collection. Ever. • Your conversations, settings and files stay in your pocket

A full-strength assistant • Streaming chat with any Ollama or OpenAI-compatible model • Plan mode and Research mode (live web search with cited sources) • Personas, conversation history, voice input and hands-free Walking Mode • Generate images right in the chat

MyLLMos — apps you imagine become apps you own Describe an app in plain words and watch the assistant build it, live, as a real mini-app you can keep. Or install ready-made ones from the App Gallery: journals, trackers, tutors, games. Your workspace has Documents, Images and Shared folders, so your apps and your assistant can work with the same files.

Power-user friendly • On-device GGUF models (Gemma, Qwen, Llama) — fully offline chat • Server terminal: pull and manage models on your Ollama box from your phone • Agent mode with tools and reusable skills • Automations that run on a schedule • Share anything into MyLLM — it lands in your Documents folder

Simple pricing Free, full-featured for your first month. Then £19.99 once — yours for life — or £0.99/month / £9.99/year. Cancel any time in Google Play.

MyLLM is for people who love AI but don't love handing their life to a cloud. If you've got an Ollama server — or just a decent phone — you've got everything you need.

Requires an Ollama or OpenAI-compatible server for the full experience; smaller open-source models run entirely on-device with no server at all.

What’s new

On-device chat that stays quick and cool. Models on your phone now pick up where they left off instead of re-reading the whole chat before every reply, so long chats stay fast. A chat that outgrows the model's memory no longer comes back blank: older messages are set aside. The model's memory takes half the space and grows only when needed. When your phone gets hot, replies slow down to let it cool. Tap the model name at the top of the chat to switch models. Downloads on mobile data ask first.