A React Native mobile app that runs language models, text-to-speech and tool use entirely on your phone, offline and without an account.
Mobile-Artificial-Intelligence/maid
Android app for chatting with GGUF models running locally via llama.cpp, or with remote providers like OpenAI, Anthropic, Ollama and DeepSeek.
Open-source alternative to:ChatGPT
- Stars
- 2,755
- Forks
- 307
- Open issues
- 11
- Last commit
- 9 days ago
What is it?
Maid is a mobile chat client for large language models on Android. It runs GGUF models directly on the device through llama.cpp, so conversations can work without an internet connection, and it can also connect to remote services such as OpenAI, Anthropic, DeepSeek, Mistral and Ollama using your own API key.
The app includes one-tap downloads of curated Hugging Face models, and you can load any GGUF file from local storage. Generation parameters like temperature, top-p, top-k and context length can be adjusted per session, and a global system prompt sets the assistant's persona.
Conversations can be created, renamed, deleted, exported and imported as JSON. The project is built with React Native, is MIT licensed, and is distributed through GitHub releases and Google Play.
Key features
- 01Runs GGUF models fully on-device via llama.cpp without internet access
- 02Connects to many remote providers with your own API key, plus LiteLLM proxies
- 03One-tap download of curated Hugging Face models inside the app
- 04Adjustable generation parameters such as temperature, top-p, top-k and context length
- 05Chat export and import as JSON
- 06Light and dark Material You theming
Use cases
- Chatting with small open models on a phone while offline
- Trying out a hosted model from OpenAI, Anthropic or DeepSeek on Android
- Using a self-hosted Ollama server as a mobile front end
- Keeping local chat history that can be backed up as JSON
Quick start
- 1
Clone the repository
git clone https://github.com/Mobile-Artificial-Intelligence/maid.git - 2
Install dependencies
yarn install - 3
Build the Android APK
yarn build-android
Things to know
- Android only; no iOS or desktop build is mentioned.
- Local models require running GGUF files on the phone, and remote providers need your own API key.
In collections
Topics
FAQ
- What is maid?
- Android app for chatting with GGUF models running locally via llama.cpp, or with remote providers like OpenAI, Anthropic, Ollama and DeepSeek. Maid is a mobile chat client for large language models on Android. It runs GGUF models directly on the device through llama.cpp, so conversations can work without an internet connection, and it can also connect to remote services such as OpenAI, Anthropic, DeepSeek, Mistral and Ollama using your own API key.
- Is maid free to use? What license does it have?
- maid is open source under the MIT license. Check the license terms before using it in commercial products.
- Is maid still maintained?
- Actively maintained. The last commit to maid was on Oct 29, 2026.
- What can I use maid for?
- Chatting with small open models on a phone while offline Trying out a hosted model from OpenAI, Anthropic or DeepSeek on Android Using a self-hosted Ollama server as a mobile front end Keeping local chat history that can be backed up as JSON
- How do I install maid?
- Clone the repository: git clone https://github.com/Mobile-Artificial-Intelligence/maid.git Install dependencies: yarn install Build the Android APK: yarn build-android
- Where can I get maid?
- The source code and installation instructions are on GitHub: https://github.com/Mobile-Artificial-Intelligence/maid
Summary generated by AI from the project's README. GitHub stats refresh daily.
Related projects
React Native binding of llama.cpp for running GGUF LLMs on-device in iOS and Android apps.
A native mobile frontend for chatting with LLMs, running gguf models on-device or connecting to commercial and open source APIs.
A Flutter chat client that runs GGUF models on-device on Android and iOS, or switches to cloud APIs from OpenAI, Anthropic, Gemini, or Kimi.
Android chat app that sends one prompt to multiple LLM providers at once, using your own API keys.
Single-file local runner for GGUF language models with a built-in KoboldAI web UI, supporting text, image, speech and music generation.