Fix multi-turn chat: lock composer during generation, always stream into a fresh end-of-history bubble

Keep header CTA clear of the floating HF Space pill

Remove stray build artifacts from repo

Add webAI app waitlist call to action; remove stray build artifacts

Extract text from Gradio 6 content blocks: model was receiving empty user turns

Enable verbose error reporting

Fix history filter dropping every message: model never saw user input

Fix blank screen during generation, stream thinking tokens, Enter to send, lower GPU duration

Add Showcase tab: one-click real-world argument analysis gallery

Fix unclickable prompt: stop hiding the textbox label, keep send on the same row

Restore the Ask webAI prompt pill that CSS had hidden

Stop the Parameters accordion from stretching to full viewport height

Keep the composer pill fixed-height and show Logicizing immediately on submit

Polish webAI UI and collapse Logic traces by default

Brand the demo like webAI and add the TwIL Intelligence Lab system prompt

Upload /tmp/twil-lm3-official-space/app.py with huggingface_hub

Add official TwIL-LM3 Gradio ZeroGPU demo

initial commit