Add upload transcript to API

[Joopsnijder]

Add a way to upload the transcript of podcast with the API, preferably in the POST /episodes.json.

[johnloringpollard]

Thanks for letting us know what you need!

We would have to discuss how to best to this, file_url? file_contents? Will it automatically write over an existing transcript?

[bbelderbos]

This would be a really nice addition. Can I help out?

[esegui]

was this created? I can't find any documentation. I would love to be able to upload Chapter Markers and SRT files via API

[johnloringpollard]

This has not been created but thank you for letting us know what would be useful for you.

[tomrossi7]

@esegui we don't have the ability to add Chapter Markers, but we did add the ability to upload transcripts. Unfortunately, we haven't had time to update the docs.

You can post the transcript to this URL:
POST /api/:podcast_id/episodes/:episode_id/transcript

You will need to pass in two parameters, the transcript_file which should be a file and the transcript_format which would be "SRT" in this case.

Give it a shot and let us know if you need more help with it!

[tomrossi7]

We're also happy to receive PRs for documentation! 🤣

[esegui]

I used the HTTP module in make.com to upload a podcast episode and it worked but to upload the SRT file, I have this error "Error: 500 Internal Server Error"

I used POST method to URL https://www.buzzsprout.com/api/xxxx/episodes/xxxx/transcript

{
"transcript_file": "https://xxx.s3.us-east-1.amazonaws.com/xxx/my.srt",
"transcript_format": "SRT"
}

Thoughts?

[esegui]

Additionally, I tried a txt file like so since we can copy and paste text when adding a transcript on the website

{
"transcript_file": "https://xxxx.s3.us-east-1.amazonaws.com/xxxx/Daily_Podcast-2024-11-127Z.txt",
"transcript_format": "txt"
}

Same 500 error

[tomrossi7]

For the transcript_file, post the actual file (not a URL).

[n8fr8]

Thanks for this feature. However, I'm having a similar challenge with the transcript upload POST, receiving this error:

transcript upload response: 422 {"errors":["cannot be parsed"]}

My code is as follows:

headers = {
"Authorization": f"Token token={api_key}",
"User-Agent": "Buzzsprout Python Uploader"
}

episode_id=responseJSON["id"]
episode_url=responseJSON["audio_url"]
transcript_file_path = audio_file_path + ".txt"

    with open(transcript_file_path, "rb") as transcript_file:
      tfiles = {
        "transcript_file": (transcript_file_path, transcript_file, "text/plain")
      }

      tdata = {
            "transcript_format": "txt"
      }
      turl = f"https://www.buzzsprout.com/api/{podcast_id}/episodes/{episode_id}/transcript"
      print("posting transcript to: " + turl)
      tResponse = requests.post(turl, headers=headers, data=tdata, files=tfiles)
      print(f"transcript upload response: {tResponse.status_code} {tResponse.text}") 

[johnloringpollard]

What is the transcript? The error indicates that it's unable to parse the transcript file.

[n8fr8]

It is just words in a text file.

[johnloringpollard]

If you put this in the txt file does it work?

Travis  0:00
Welcome back to buzz cast. We got a pretty cool episode for you got a very special guest

Karin joining us on the episode today. Hello, Karen.

Kerry-Anne  0:11
Hello, how are you?

And this should work for an SRT file.

1
00:00:00,000 --> 00:00:04,020
Alban: Can I give you one surefire
way to get removed from Spotify?

2
00:00:04,320 --> 00:00:04,890
That always works?

[n8fr8]

adding the timestamps to the txt format didn't change anything.

Will try generating an SRT file instead, and let you know.

Thanks!

[lucpelletier]

@n8fr8 Did you ever manage to get this to work?

I have tried a number of different formats and methods and I always get an error (either 422 or 500).

I used POST to send a json object like the following and it fails for both VTT and SRT formats:

{
  transcript_format: "VTT",
  transcript_file: "WEBVTT\\n\\n00:00:00.000 --> 00:00:19.360\\n<rest of vtt file>"
}

transcript_file contains the whole VTT file but I've cut it short here just to give an idea of what it looks like.

Can one of the developers please clarify what the request body should contain exactly? Is it json or form-data? Which transcript formats are supported?

Thanks in advance.