Ratings and reviews

This works great, and seems easy to use.

ottima …

Worked for me on one of my videos. It only misspelled two words, both I wouldn’t have expected it to get correct. And I found it to be accurate for my video. And I do appreciate that there is an option to globally change the spelling of a word and love the fact that I can use it to output captions for the video. And even better I can create multi-language translations of the captions. Truly a great help. This is on a M5 Macbook. I was impressed.

Perfect! It transcribes also TikTok, YouTube and Instagram REELs. I am using M2 PRO, 16GB RAM.

无法下载模型，下载一半卡住不动了

中文无法识别，导出空白

Simply not accurate. When I put an audio file through a standard LLM (e.g. ChatGPT, Claude) I get a transcript that is far, far more accurate than what I get from this tool. Not worth using. Edit: The developer has responded and I feel compelled to offer more detail given their rather defensive response. I am fully aware of the advantages (and potential drawbacks) of an online solution like those that I mentioned. But it isn’t that they are a bit more accurate. I shared a crystal clear recording of a meeting with this app using the LLM option labaled most accurate, and it was practically nonsensical. There is gibberish, and at times what appears to be German translated (despite explicitly selecting English). Nice to know that the developer “respects my opnion” but respectfully, this isn’t an opinion. This app doesn’t work.

Keeps crashing

Até o momento o melhor App de transcrição que conheço!

I tested more than one hundred files in Spanish and English and obtained very good transcriptions. Parakeet is really fast and produces good transcripts. Finally I was able to transcript many audios of courses I did some years ago. Some weird things I found: - Memory is not being freed when a transcription is finished. The app ended up using more than 42 GB of RAM and caused my laptop to freeze multiple times. When bulk exporting, I had to select smaller folders and close and reopen the app before starting another bulk export. - At the beginning I had two models downloaded: Whisper and Parakeet. When I bulk exported a folder, all the transcripts were generated using Whisper even though I had selected Parakeet as the transcription model. I had to leave only one model to avoid this error. Thanks for this great tool, it is very useful.

This is exactly what setapp needed, i love it

So far it looks nice, but very rough. I'll update this as I have more experience with it. Here are the problems I've found so far: - The dialog box to request access for screen recording appeared before the "Welcome to TranscribeX" modal, and then was covered by the welcome modal. This might not be a problem except that the welcome modal could not be moved independently, which seems like a slightly user-hostile decision. First: avoid covering up your own dialog boxes. Second: allow users to move them wherever they need to. Don't prevent us from performing basic operations like that. Next, please add support for GitHub Copilot LLMs! Lastly, the automatically assigned colors for speakers were very repetetive and didn't avoid giving two people the same color. That was made worse by the color picker being very buggy for me, changing the color assignment in ways contrary to what I was telling it to do, but that might be Apple's fault.