feat(voice-to-text): default files and recordings to Whisper; live dictation always uses Fast

[slaveofcode]

• Files and recordings now default to the Multilingual (Whisper) model, which is more accurate. Fast is still one click away, and the Whisper models are listed first.
• Live dictation always runs on the Fast engine, whatever model is picked. It's available whenever the language is one of Fast's 7 languages or Auto-detect.
• The live progress label checks the Fast engine's cache separately, so it no longer reads the cache of the selected file model.
• The UI hints, EN + ID SEO copy (intro, how-to, model FAQ) and the beta.7 CHANGELOG line are updated.
• E2E: the default is Whisper and Live is enabled; the Fast-to-Whisper switch for Indonesian still works; the file test picks Fast explicitly. All 4 tests pass locally.

[cloudflare-workers-and-pages]

Deploying with    Cloudflare Workers

The latest updates on your project. Learn more about integrating Git with Workers.

Status	Name	Latest Commit	Updated (UTC)
✅ Deployment successful! View logs	goodwebtools	247f004	Oct 12 2026, 02:35 AM