#Transcribe audio files to text
30 messages · Page 1 of 1 (latest)
@vivid hull Are songs included?
@karmic willow Songs are included!
Yay, tried it on English songs with 70% accuracy
which programming languages did you write? 🙂
No plans to open source but I’m using https://github.com/openai-php/client as my main method to interact with the whisper api.
PHP is the backend language
@vivid hull Proposing some fixes to the application:
Fix the upload size that it allows up to 25MB audio files in one file (increased capacity). Or correct the error message to 12MB and restrict uploading file to <12MB: "The upload field must not be greater than 12288 kilobytes."
Make the text space wider for computer users.
And inform uploaders if their data would be saved in a storage or removed after processing, as some users are particular with file privacy.
Make the "Server Error" message box smaller, and add an X button or "Okay"
I appreciate the ideas! I will work on getting them implemented.
When uploading a file, i inmediatly get the "500 server error"
The 500 error was most likely because of a max upload error. This has been resolved. If you attempt a file larger than 12MB but less than 25MB, it should generate the transcript now.
Better file handling post processing has been implemented. There is an opt in checkbox that would allow for uploaded audio to be used for testing and improvement purposes. If the checkbox is not checked, the file is immediately removed after it is processed.
Does it also support korean
it should support the language
I have added the ability to insert a TikTok or YouTube url and get a transcription back. YouTube is still hit and miss depending on the size of the video. Let me know what you think
https://transcribeai.io/free/video-transcribe
@vivid hull , good work. I will try it on url links. Do u have api, people could use
Just tried with YouTube. It’s not giving results. Tried TikTok. It also failing. Does it work for any video link like Instagram. TikTok , YouTube. It shouldn’t matter right , If u r using whisper; it converts audio to text.
I have found a good way to split up audio files over 25MB which Youtube videos tend to be. I should have a better method in place soon. Whisper doesn't care as long as it's an audio file. The hardest part is getting the audio file from various outlets. I'll post when I have the new method in place.
Twas a miss! (Original reference to the video is redacted)
A workaround is downloading the video first, then uploading it to the site
But in another, it worked
I’ll take a look into the issue.
Follow-up for @vivid hull