Whistle: Speech to Text in 16.9 MB | Hacker News - But Better

Whistle: Speech to Text in 16.9 MB cactuscompute.com

An open speech recognition model that runs on the same CPU engine as Needle. It transcribes seven languages, reaches the first token in 11 ms, and loads beside Needle so one binary turns a clip straight into tool calls.

1 points•yarapavan•about 3 hours ago•0 comments•

#speech #whistle #text #needle #cpu #recognition #engine #transcribes

0 comments

No comments yet.

Related stories

Whistle: Speech to Text in 16.9 MB by Cactus Computecactuscompute.com
Hacker News · 1 points · about 14 hours ago
Show HN: Universal 3.6 Pro – streaming speech-to-text tuned for voice agentsassemblyai.com
Hacker News · 1 points · 4 days ago
Rai: CPU-only LLM inference engine in pure Rustgithub.com
Hacker News · 2 points · about 18 hours ago
Show HN: Cactus Needle 3: 8-29MB automation models can match DeepSeek V4 Flashcactuscompute.com
Hacker News · 236 points · 15 days ago
Oído: Open-vocabulary speech recognition on a $5 ESP32-S3 (3.7% LibriSpeech WER)github.com
Hacker News · 2 points · 3 days ago
Speech-to-Text Accuracy in 2026: Open-Weight vs. Cloudverli.app
Hacker News · 2 points · 6 days ago