ilmora
Private by design

Your AI.
Your device.

Think, write, and explore with language models that run on your Android phone. No account, no cloud chat history, no lock-in.

Get it on Google Play ilmora - Your AI. Your device. | Product Hunt
See how it works
  • Local inference
  • GGUF models
  • No account required
0accounts required
Localconversation storage
GGUFmodel flexibility
Androidbuilt for mobile
Calm, capable, private

Everything you need.
Nothing you don’t.

ilmora keeps the interface out of your way while giving you control over the model, conversation, and data.

Private conversations

Prompts, responses, and session history are processed and stored on your device for core chat.

Local modelNo cloud chat backend

Bring your own model

Import a GGUF file from your device and switch models without changing how you chat.

LlamaQwenPhiGemma

Speak naturally

Optional voice dictation uses the speech recognition service installed on your phone, including Google Speech Services where available.

Conversations that stay organized

Create, rename, search, export, and restore sessions. Give each conversation its own assistant instructions.

Smoother on mobile

Memory-safe model imports and frame-aware token streaming keep the interface responsive during demanding local work.

Three simple steps

From model file
to first thought.

You stay in control from setup to every conversation.

  1. 01

    Add a model

    Choose a compatible GGUF file already saved on your Android device.

  2. 02

    Pick your assistant

    ilmora detects the model family and applies the matching chat format.

  3. 03

    Type or speak

    Start a focused conversation, with streaming replies and simple stop controls.

Designed around ownership

Your thoughts are not a data product.

ilmora has no account system, ads, analytics SDK, or cloud chat database. Core model inference stays on your device.

Read the privacy details
Good to know

Common questions

Does ilmora need internet access?

Core model inference and chat storage do not. Optional voice recognition may use your device’s configured provider and a network connection. You source model files separately.

Which models work?

ilmora supports GGUF models and recognizes common Llama, Mistral, Qwen, Phi, Gemma, and related chat templates. Smaller 1B–3B quantized models are usually the best fit for phones.

Where are conversations stored?

Conversation data and settings are kept in the app’s local storage. You can delete sessions or export them as JSON.

Think locally.

Download Ilmora and make private, on-device AI part of your everyday thinking.

Get Ilmora on Google Play