Standalone Chatbot Agent

ℹ️ About This Demo

This demo showcases running AI models directly in your browser. You can switch between:

• SmolLM2-135M-Instruct: A compact language model from Hugging Face (works without token)
• SmolLM-360M-Instruct: Larger SmolLM model with better performance (works without token)
• Nikity/lille-130m-instruct: Custom tokenizer model (currently not working)
• Gemma-3-270M-IT: Google's Gemma 3 model (requires Hugging Face access token)
• Transformers.js: Running models client-side with WebAssembly and ONNX runtime
• Local Processing: All AI computation happens in your browser - no data sent to servers
• Hardware Acceleration: WebGPU support for faster processing when available

Note: Some models require a Hugging Face access token for access. Use the token input section below to provide your token.

This demonstrates modern web AI capabilities, running language models entirely in the browser.