100% Local
Privacy First
All AI inference happens locally in your browser. No data sent to external servers.
Now running on Edge
AI that runs
where you are
Run Falcon-H1, Whisper, and Kokoro AI models directly in your browser. No servers. No data collection. Just privacy-first AI.
4 Models Active
~250MB Total Size100% Browser-BasedFeatures
Works Offline
Offline Capable
Models are cached locally. Works without internet after initial setup.
GPL 3.0
Open Source
GPL 3.0 licensed. View, modify, and distribute the code freely.
Edge Ready
Edge Deployment
Runs on Cloudflare Workers, Vercel Edge, and any edge runtime.
Powered by Open Models
All models are sourced from HuggingFace and optimized for edge deployment. Click any model to view its source.
Chat AI
Falcon-H1-Tiny-90M
Language Model
90M
parameters
Instruction-tuned chat model for conversational AI. Q4_F16 quantized for browser.
~91MB
HuggingFaceSTT
Whisper Tiny
Speech Recognition
39M
parameters
End-to-end speech recognition for transcription and voice commands.
~31MB
HuggingFaceTTS
Kokoro-82M
Text-to-Speech
82M
parameters
High-quality neural voice synthesis with multiple voice options.
~82MB
HuggingFaceEmbeddings
all-MiniLM-L6-v2
Embeddings
22M
parameters
Sentence transformer for semantic search and RAG applications.
~45MB
HuggingFaceView all model source URLs
- Falcon-H1-Tiny-90M
- https://huggingface.co/onnx-community/Falcon-H1-Tiny-90M-Instruct-ONNX
- Whisper Tiny
- https://huggingface.co/Xenova/whisper-tiny
- Web Speech API
- Web Speech API
- all-MiniLM-L6-v2
- https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2
100%
Client-Side
4
AI Models
~250MB
Total Size
GPL 3.0
License
Ready to try it yourself?
Open the playground and chat with Falcon-H1 directly in your browser. No account required.
Start Chatting