This thread is a support group for people who have each independently built the same macOS speech-to-text app.
https://opensource.builders/alternatives/superwhisper
Just added Ghost Pepper, and you can actually create a skill.md with the features you need to build your own
Wish they would do an ios version, but the creator already kind of dismissed it.
I didn't just dismiss for no reason, I am a human! I have needs and I can't just sleeplessly stay in front of the computer putting out code. If I had more time I would, but alas.
Someone could easily vibe code an iOS version in a few hours. I could do the same but I do not have time to support it.
https://apps.apple.com/us/app/voiceink-ai-dictation/id675143...
(I was searching the same as you before I found this last month)
It has all the usual features, plus you can add project specific vocabulary in your repo. It detects the working folder based on the active window, reads a WORDBIRD.md file in that folder and corrects terms accordingly.
(My friend Till built it)
1: livestream transcript directly into the cursor in real time (just like native macOS dictation)
2: show realtime transcript live in an overlay (still has to paste when done, unlike #1, but can still read live while dictating)
1- localvoxtral, 2- FluidVoice (bumping it to 7 features on your list)
I see promise trying to get a bit more into curating by showing the top one or two or three picks for a given standout feature.
Like telling it to edit the text or remove a word.
The cherry on top: it’s completely broken! Enable the Context Awareness filter, the list shrinks. Now enable the Auto-pasting filter, the list grows back.
The categories are clearly LLM generated from the GhostPepper codebase, with vague low level descriptions and links to code. Most categories apply to every listed project.
The UI is the same tiny bit of LLM generated information displayed five different confusing ways. Like seriously, click on a project and you first see a bunch of haphazard feature cards, then a bunch of “feature ... active” rows. Looks fancy, but actually just noise. Textbook slop.
Better would be a simple awesome-style markdown page, with a feature matrix having categories and descriptions curated by a human that actually understands and cares about the domain.
Sorry if this is harsh, but passing off LLM output as “curation” is particularly insulting to me.
>“Compare” - This is the most important part. Apps in the most saturated categories (whisper dictation, clipboard managers, wallpaper apps, etc.) must clearly explain their differentiation from existing solutions.
https://www.reddit.com/r/macapps/comments/1r6d06r/new_post_r...
That whole list of requirements there is actually a good thing that anyone who wants to make a new application should ask themselves.
It's remarkable how similar its performance is to Wispr Flow... and it runs locally...
When I most recently abandoned it, the trigger word would fire one time in five.
I built one for cross platform — using parakeet mlx or faster whisper. :)
But I did it because I wanted it to work exactly the way I wanted it.
Also, for kicks, I (codex) ported it to Linux. But because my Linux laptop isn't as fast, I've had to use a few tricks to make it fast. https://github.com/obra/pepper-x
This is the unfortunate real face of open source. So many devs each making little sandcastles on their own when if efforts were combined, we could have had something truly solid and sustainable, instead of a litany of 90% there apps each missing something or the other, leaving people ending up using WisprFlow etc.
windows (kotlin multi platform) => https://github.com/maceip/daydream
parakeet-tdt-0.6b-v2
hotword dict so no more "clawd" "dash" "dot com"