🛠️ Overtone – Transcribe, correct and cut captions for any screen shape
Summary
Overtone is an open-source tool that transcribes audio using multiple models and an LLM to reduce transcription errors. It also includes a rules-based cutter to format captions for any screen shape.
Why it’s interesting
It uses multiple speech-to-text models combined with an LLM sanity check to lower the Word Error Rate and automates caption layout formatting.
Source metrics: Points 13 · Comments 0
HN discussion · Project
Source: #HackerNews / Show HN
Summary
Overtone is an open-source tool that transcribes audio using multiple models and an LLM to reduce transcription errors. It also includes a rules-based cutter to format captions for any screen shape.
Why it’s interesting
It uses multiple speech-to-text models combined with an LLM sanity check to lower the Word Error Rate and automates caption layout formatting.
Source metrics: Points 13 · Comments 0
HN discussion · Project
Source: #HackerNews / Show HN