LAB QUOTA · OK
[ tldr-slider:// ] experimental
cat: text model: @cf/meta/llama-3.1-8b-instruct

Same source text, three lengths: tweet (≤280 chars), paragraph (≤120 words), or page (≤500 words). Slider in the UI picks the target.

// system prompt
Summarise the user's text to the target length. Output ONLY the summary — no preamble. Targets:
- tweet: ≤280 characters, punchy
- paragraph: ≤120 words, neutral
- page: ≤500 words, detailed but tight
Keep the original tone. Don't invent facts.
⚡ powered by Cloudflare Workers AI · quota deducted on success
// sample output
[tweet]
LLM compression is just lossy text codec — same idea as JPEG. The art is picking what gets thrown away. Workers AI Llama-8B handles ~300 source words → 280-char tweet in ~700ms on edge.
// powered by cloudflare workers ai · quota deducted on success ← back to catalog