Thank you for the tutorial! Have you tried to do it as real-time streaming (ASR)? I'm stuck with NodeJS doing that. Seems like OpenAI API works only when the media recorder stops. I try to pass the audio chunks every 2 seconds, the first request succeed but the subsequent ones fail.
hello, thank you for the video, it's really very userful. But maybe you can help - how can I specify the language of recognition, for example I need the Czech language to be recognized on whisperAI? Thank you very much for your help
Hello, thank you very much for the video. It has been very useful to me and I have learned a lot. After carrying out some tests, in local mode it works very well but if the tests are in a Hosting, a Json error appears in the console. ---- Sending audio file to server... input.js:65 Error sending audio file to server: SyntaxError: Unexpected token 'S', "Sorry, the"... is not valid JSON at JSON.parse () at input.js:59:44 (anonymous) @input.js:65 ---- Do you know what it could be? Thank you so much
Same problem here. Error sending audio file to server: SyntaxError: Unexpected token 'S', "Sorry, the"... is not valid JSON at JSON.parse () Did you find out what the problem was?
To solve this you may first change some of the "echo" of the whisper_send_data.php to some logging function, to understand what goes under. You may then set CURLOPT_SSL_VERIFYPEER to false (except if you are unsure about openai URL securization) ; then you'll learn on the Internets that the CURLfile constructor sometimes needs to have a canonical path, so you will change $target_file to realpath($target_file). Thanks to the initial author for these scripts !
![[UNCUT] "บอสณวัฒน์" แฉ! แหล่งซุกเงินแก๊งบอส ขยันผิดที่ไม่กี่ปีก็เข้าคุก l คนดังนั่งเคลียร์ l17ต.ค.67](http://i.ytimg.com/vi/BL56qIfXYZM/mqdefault.jpg)
![เปิดใจ "อาจารย์เบียร์ คนตื่นธรรม" อดีตเคยสุดทุกทาง ถ้าใช้หนี้หมดไปบวชแน่ l [Nickynachat]](http://i.ytimg.com/vi/IxeE0xVhy08/mqdefault.jpg)
That's awesome. I was looking for something in js and php ! Thanks
Glad it helped!
Thanks a lot for the video and the code. Helped me a lot !
Glad it helped!
Thank you for the tutorial!
Have you tried to do it as real-time streaming (ASR)? I'm stuck with NodeJS doing that. Seems like OpenAI API works only when the media recorder stops. I try to pass the audio chunks every 2 seconds, the first request succeed but the subsequent ones fail.
You saved my day! Thank you very much!
Anytime
Nicely done. Explained well
Glad it was helpful!
You’ve been the best so far
But please
Please can you show how to use google api for ocr through ph
Thank you for the suggestion, I will work on one! Contact me directly if you need help at InteractiveUtopia.com
hello, thank you for the video, it's really very userful. But maybe you can help - how can I specify the language of recognition, for example I need the Czech language to be recognized on whisperAI? Thank you very much for your help
You can do this by telling the languaje the input is in, look at platform.openai.com/docs/api-reference/audio/createTranscription
@@InteractiveUtopia thank you!
hello from Chile :) thanks for your video. im doing the test but i need to have a plan on open ia to do the test? "insufficient_quota" problem :/
Hello, thank you very much for the video. It has been very useful to me and I have learned a lot. After carrying out some tests, in local mode it works very well but if the tests are in a Hosting, a Json error appears in the console.
----
Sending audio file to server...
input.js:65 Error sending audio file to server: SyntaxError: Unexpected token 'S', "Sorry, the"... is not valid JSON
at JSON.parse ()
at input.js:59:44
(anonymous) @input.js:65
----
Do you know what it could be?
Thank you so much
Same problem here.
Error sending audio file to server: SyntaxError: Unexpected token 'S', "Sorry, the"... is not valid JSON
at JSON.parse ()
Did you find out what the problem was?
probably need to make sure you change the OPENAPI_KEY.....worked for me.@@ramon6456
To solve this you may first change some of the "echo" of the whisper_send_data.php to some logging function, to understand what goes under. You may then set CURLOPT_SSL_VERIFYPEER to false (except if you are unsure about openai URL securization) ; then you'll learn on the Internets that the CURLfile constructor sometimes needs to have a canonical path, so you will change $target_file to realpath($target_file).
Thanks to the initial author for these scripts !
Thank you !