Build Small Hackathon · 15.5B parameters total
Split-Brain Co-Pilot
One small model drafts in your browser. Another small model checks it on Modal. The UI shows the handoff, verdict, rollback, and executable proof.
WebGPU local-first
llama.cpp verifier
Modal sandbox
Custom Gradio UI
Space initializing...
Language
Local Draft
WebGPU 1.5B
fast browser stream
Cloud Check
Modal A10G 14B
llama.cpp verifier
WebGPU not detected. Use Chrome 113+ on desktop for local inference.
Model not loaded
Speculative draft
Idle
Waiting for model load...
Idle
0 tok/s
Verifier idle