Offline AI Chat
Download a local LLM once, and enjoy smart, natural conversations anytime—even without internet access.
Built for privacy, speed, and flexibility
Privacy First
Your data never leaves your device. No tracking, no servers. Just you and your private AI assistant.
Custom Models
Choose from optimized open-source models designed for mobile performance and accuracy.
Local API
Access your device and local LLM over your local network with a secure API. Data stays on your device and LAN.
Available Models
Llama
Qwen3
Deepseek
Gemma
FastVLM
SmolVLM2
Stable Diffusion 1.5
...and more
Abliterated models
Abliterated models are large language models (LLMs) that have been modified to remove or reduce their built-in refusal mechanisms, allowing them to generate responses they would normally reject. This is achieved by a technique called "abliteration," which involves analyzing and manipulating the model's internal workings to bypass its safety filters.
All Abliterated models in AIBro are designed to provide safe and legal content only.
MNN engine
MNN is a blazing fast, lightweight deep learning framework, battle-tested by business-critical use cases. AIBro is based on MNN Chat with new unique models.
API Access
Use the local API to connect apps on your network to the LLM running on your device. Your data stays within your local network.
GET http://DEVICE_IP:PORT/v1/chat
{
"model": "llama-mobile-v1",
"messages": [{"role":"user","content":"Hello"}]
}
Take control of your AI experience
Download now and chat freely, anywhere, anytime — offline and secure.
