https://avilpage.com/enFri, 31 Jul 2026 17:07:15 GMTNikola (getnikola.com)http://blogs.law.harvard.edu/tech/rsshttps://avilpage.com/2026/07/handy-stt-dictation-cross-platform.htmlAnand Reddy Pandikunta<h4>Context</h4> <p>Around 2016, I explored various speech-to-text (STT) tools to reduce strain on writs due to RSI(Repetetive Strain Injury) and gave up on them due to their limitations.</p> <p>With AI advancements, these tools have come a long way.</p> <h4>Open Source STT Tools</h4> <p><a href="https://github.com/cjpais/Handy">Handy</a></p> <p><a href="https://github.com/TypeWhisper">TypeWhisper</a></p> <p><a href="https://github.com/altic-dev/FluidVoice">FluidVoice</a></p> <p><a href="https://github.com/ray-amjad/hyperwhisper-app">HyperWhisper</a></p> <p><a href="https://github.com/Beingpax/VoiceInk">VoiceInk</a></p> <p>There are couple of paid tools but they didn't work well in noisy environment.</p> <p>After trying out all these tools with local models like Parakeet, Voxstral, and Cohere Transcribe, I found that Handy + Cohere Transcribe is the best combination for my needs.</p> <p><img alt="mac-tts-models.png" src="https://avilpage.com/images/mac-tts-models.png"></p> <p>Handy has post processing as an experimental feature where we can send output of STT to other LLM models and get a fine grained output.</p> <p><img alt="mac-tts-post-process.png" src="https://avilpage.com/images/mac-tts-post-process.png"></p> <p>For example, when I dictate "git push dash dash force", the output is "git push --force" after post processing.</p> <h4>Conclusion</h4> <p>I let AI to write <a href="https://github.com/ChillarAnand/init/blob/main/handy_stats.py">handy_stats.py</a> script to track my dictation stats. I am dictating more than 10K characters per day which reduces strain on my fingers/writs.</p>cross-platformdictationspeech-to-texthttps://avilpage.com/2026/07/handy-stt-dictation-cross-platform.htmlTue, 28 Jul 2026 23:49:39 GMT