About Apollyon
Apollyon is a cutting-edge AI assistant that runs entirely in your browser. Your conversations stay private and secure on your device, with no data ever being sent to external servers.
Key Features
- Complete privacy - all processing happens locally in your browser
- No internet connection required after initial model download
- Runs on the CPU via WebAssembly - no WebGPU or dedicated graphics card needed, so it works on far more browsers and devices
- You running the model means no ratelimiting whatsoever.
- Fast and responsive interactions with streaming responses
- Multiple chat management with automatic AI-generated titles
- Export your conversations for future reference
Getting Started
Simply wait for the model to load, then start typing your questions or requests. The AI will respond in real-time as it generates its answer. You can create new chats, switch between conversations, and export your chat history at any time.
Technical Details
Powered by the Llama 3.2 1B Instruct model in GGUF format, run through wllama, a WebAssembly binding for llama.cpp. Inference happens entirely on the CPU, so no WebGPU support is required - this trades some raw speed for compatibility with essentially any modern browser, including ones without GPU acceleration. Response speed depends on your device's CPU; more cores generally means faster generation.