Hi, I tried searching but only found Text-to-Speech alternatives.
I’m looking to record a lot of text and was wondering about the best way to achieve Speech-to-Text with privacy respecting software.
Thank you!
Hi, I tried searching but only found Text-to-Speech alternatives.
I’m looking to record a lot of text and was wondering about the best way to achieve Speech-to-Text with privacy respecting software.
Thank you!
I haven’t checked but the proliferation of generated AI YouTube channels, you should be able to host your own LLM and generate TTS somehow…
I want to adopt a transcription tool, but unsure on what is the best one for ease of use + privacy. What are folks using these days / best when it comes to privacy guarantees?
On my self-hosted Open WebUI instance, it uses Faster Whisper locally:
thanks for sharing, what pushed you to use faster whisper? any problems with it so far?
It is simply the default STT option. I do not use it often, but it carries the same problems as other transcription software, you need a good microphone, clear pronounciation, controlled room ambiance, etc.
I’ll check it out, thanks for the tip!
how has it been so far?
Framasoft developed an apl for transcription called Lokas. lokas.app
For privacy, I’d keep the first filter simple: does the audio leave your device? Local Whisper options are usually the safer starting point than a polished cloud service, because raw voice recordings and transcripts can contain a lot more personal detail than people expect. If setup effort matters, I’d try a small local Whisper app first, then only move to something self-hosted like faster-whisper if you need batches or a web UI.