Realtime API: [transcription failed] using Playground mode

API Bugs
1

I’m trying to test Realtime API on Playground mode, but every time I record some audio, it returns “[transcription failed]”. The logs contain:

type:"conversation.item.input_audio_transcription.failed"
event_id:"event_AEeq7dSSsMP7eKPvUuAoi"
item_id:"item_AEeq2FVAzno5jIkd245Jk"
content_index:0
error.type:"server_error"
error.code:null
error.message:"Input transcription failed for item 'item_AEeq2FVAzno5jIkd245Jk'."
error.param:null

I tried with different mics, saying different things, and it only returns [transcription failed] or inaudible.

Is anyone else having this problem too?

I’m not sure if this is an instability, or a bug, or if I’m doing something wrong.

Captura de Tela 2024-10-04 às 12.57.19
Captura de Tela 2024-10-04 às 12.57.191087×1184 117 KB

1 Like
2

same issue here. looks like that even though it does show on the Playground, I still don’t have access to the API itself.

Screenshot 2024-10-04 at 18.43.23
Screenshot 2024-10-04 at 18.43.231064×588 68.5 KB

3 Likes
3

For me it was working fine for a couple hours and then suddenly stopped. Is there a service crashed for transcription?

4

Hey all, thank you for the report! We identified the source of this access issue and deployed a fix. The playground should now be correctly accessible.

If you have any further issues, please report them here!

2 Likes
5

I still get the same error.

I am getting this error for the past 7 hours approx

Screenshot 2024-10-05 at 11.07.34
Screenshot 2024-10-05 at 11.07.341607×894 126 KB

It was working for me previously.
As per my dashboard I have made 109 requests in total (Tier 4)

6

You can check here and also could subscribe to service downtime emails…

7

@Praison Can you share any other details about the issue you’re seeing? Are you noticing any errors displayed near to top of the screen? We send toast error notifications for this type of error now.

Also, can you show the response.done log just below this transcription failure log to clarify it’s the same model access error?

8

@tylersmith

Screenshot 2024-10-05 at 21.28.00
Screenshot 2024-10-05 at 21.28.001578×896 161 KB

Now I see the error, rate limit exceeded

response.status_details.error.code:"rate_limit_exceeded"
response.status_details.error.message:"You exceeded your current quota, please check your plan and billing details. For more information on this error, read the docs: https://platform.openai.com/docs/guides/error-codes/api-errors."

I am in Tier 4 . Current Usage: GPT-4o-realtime-preview-2024-10-01 API requests 99 Tokens 37,270

Is there any chance I could get this rate limit increased please?
That would be great.

As I am working in an important project right now, and also will be used to create YouTube tutorials on how to use this API, showing how to integrate it in an application

Help would be much appreciated.

Thanks

1 Like
9

Hey @Praison, just dropped you a DM. Excited to check out your videos - thank you for sharing!

1 Like
11

@tylersmith Ops, such a basic issue and I was not able to identify.
Sorry for that . Now its working :slight_smile:

12

Transcription is failing for me :frowning:—both in the Playground and realtime console demo.

13

I am having the same transcription error and did not change anything in my code, using the real-time API.

14

@tylersmith @Praison
Error “conversation.item.input_audio_transcription.failed” on OpenAI Realtime Console

I wanted to follow up on the issue I encountered, which seems similar to the error others have reported above: “conversation.item.input_audio_transcription.failed” in the OpenAI Realtime Console.

when I attempt to provide voice input in both manual and VAD modes, the system does not respond as expected.
Upon checking the server response in events console, I received the following error:

server conversation.item.input_audio_transcription.failed
{
  "type": "conversation.item.input_audio_transcription.failed",
  "event_id": "event_AIB6nVHr1MUND230KbLSo",
  "item_id": "item_AIB6nf7Ym3bOulC9DKYRB",
  "content_index": 0,
  "error": {
    "type": "server_error",
    "code": null,
    "message": "Input transcription failed for item 'item_AIB6nf7Ym3bOulC9DKYRB'.",
    "param": null
  }
}

server response.done
{
  "type": "response.done",
  "event_id": "event_AIB6n7imLzijzysdtFySO",
  "response": {
    "object": "realtime.response",
    "id": "resp_AIB6novOnk6iJAiRo9TFO",
    "status": "failed",
    "status_details": {
      "type": "failed",
      "error": {
        "type": "server_error",
        "code": null,
        "message": "The server had an error while processing your request. Sorry about that! You can retry your request, or contact us through our help center at help.openai.com if the error persists. (Please include the session ID sess_AIB6k7Q9LvJd7IMN8pUeV in your message.)"
      }
    },
    "output": [],
    "usage": {
      "total_tokens": 0,
      "input_tokens": 0,
      "output_tokens": 0,
      "input_token_details": {
        "cached_tokens": 0,
        "text_tokens": 0,
        "audio_tokens": 0
      },
      "output_token_details": {
        "text_tokens": 0,
        "audio_tokens": 0
      }
    }
  }
}

Could you please assist with troubleshooting this problem?

15

Me too, transcription fails in the console app, and in playground, realtime voice is working, and I have the Whisper-1 model also enabled in my models list?

16

We had this error when our account was out of credits. After purchasing more credits the error went away. It would be nice if the playground interface gave a better error to indicate that as the cause.

17

Same problem. I’m trying to use OpenAI Realtime API Console. The only response I ever got from the server is “conversation.item.input_audio_transcription.failed”

18

Thanks a lot fro documenting this, it turned out I had run out of credits too. After topping up, the problem went away.