Ollama Mobile is the native iPhone and Android app for private AI chat. Use local network or Ollama Cloud endpoints, reuse prompts, generate images, and keep conversations on the device.
Most apps charge you, track you, or lock features behind a paywall. We don't. AI should be for everyone. Just connect to your Ollama
Ollama is one of the best tools for running open source AI models but presently there is no mobile client. Away from your desk means away from your models. We built Ollama Mobile to close that gap. Ollama mobile is a native iPhone and Android app that works with Ollama endpoints.
We follow Ollama's own philosophy. Chats, prompts, and API keys never touch our servers. The app talks directly to your endpoint and nothing is stored, logged, or routed elsewhere. Your data, your rules, exactly as Ollama intends.
A mobile AI app that makes the workflow obvious.
Connect to a local network model or Ollama Cloud. Endpoint setup is visible and explicit.
Use builtin prompts, create your own custom prompt, then jump into common jobs like summarize, explain like I am 5, or generate an image.
Chats, prompts, and keys stay local. The app is built around device storage and keychain-backed credentials, so privacy is part of the product behavior.
Sign up at ollama.com, generate a free API key, and you're ready to chat from anywhere.
Explore what you can achieve with the app.
Local network or Ollama Cloud. Endpoint setup is visible and explicit with no hidden calls.
Chats, prompts, and keys never leave the phone unless you point it at an endpoint that sends them.
Use any model on local network or Ollama Cloud. Chat history stays on the device.
Point the app at a local image generation model and create without any cloud dependency.
Light and dark modes that follow your system settings. No manual switching needed.
Summarize, explain like I am 5, generate images, ready immediately, no setup.
Write and save custom prompts for the jobs you run often.
Up and running in minutes. Connect your endpoint, pick a model, and start chatting.
Yes. The app ships with matching light and dark themes. Every chat, prompt, and settings screen renders cleanly in both, so the design language feels deliberate instead of like a skin swap.
Yes. You can ask the assistant to summarize a long thread, explain something like I am five, or generate an image without leaving the chat workflow. The model running on your configured endpoint handles the work; the app just presents the result.
Builtin prompts are one-tap shortcuts for the jobs you repeat, such as summarize or plan a trip. Custom prompts let you save your own instructions and system prompts so the assistant behaves the way you want every time. Both live in the Prompts screen and are one tap away from any chat.
Yes. The Connections and Models screen lets you add, edit, and switch between endpoints. Point it at a model running on your local network, a self-hosted server, or Ollama Cloud. Each endpoint keeps its own configuration and key.
Chats, prompts, and preferences are stored locally on the device. API keys are kept in the operating system credential store (Keychain on iOS, EncryptedSharedPreferences on Android). Nothing is synced to the cloud and there is no account required to use the app.
Use the Support page. The contact form goes to the team, and it helps to include your endpoint type, the model in use, and what you expected versus what happened so we can reproduce quickly.
