Transcribe Audio

Pipe directly integrates with Replicate to provide audio transcription using a very fast variant of OpenAI’s Whisper model Large V3. The transcription is done as part of the recording processing pipeline, and the transcript will be available as a separate JSON file. The filename will be STREAMNAME_transcript.json.

Enabling it is very easy:

1. Go to the Transcoding section in the Pipe Account Dashboard.
2. Scroll to the Audio Transcription section and toogle on Transcribe new recordings.
3. Save.

The recording audio will be sent to the Replicate servers for transcription. Audio extraction does not need to be enabled for this to work, but if enabled, it will provide the audio data for transcription.

The audio data will be kept by Replicate while the transcription runs and will be deleted immediately after the transcription is completed. The transcription data will remain on the Replicate servers for up to 1 hour after transcription is completed.

Transcription generally takes a few seconds to a few minutes depending on the length of the recording.

Language is automatically detected with the following supported languages: afrikaans, albanian, amharic, arabic, armenian, assamese, azerbaijani, bashkir, basque, belarusian, bengali, bosnian, breton, bulgarian, cantonese, catalan, chinese, croatian, czech, danish, dutch, english, estonian, faroese, finnish, french, galician, georgian, german, greek, gujarati, haitian creole, hausa, hawaiian, hebrew, hindi, hungarian, icelandic, indonesian, italian, japanese, javanese, kannada, kazakh, khmer, korean, lao, latin, latvian, lingala, lithuanian, luxembourgish, macedonian, malagasy, malay, malayalam, maltese, maori, marathi, mongolian, myanmar, nepali, norwegian, nynorsk, occitan, pashto, persian, polish, portuguese, punjabi, romanian, russian, sanskrit, serbian, shona, sindhi, sinhala, slovak, slovenian, somali, spanish, sundanese, swahili, swedish, tagalog, tajik, tamil, tatar, telugu, thai, tibetan, turkish, turkmen, ukrainian, urdu, uzbek, vietnamese, welsh, yiddish, yoruba.

The resulting transcription file will be pushed to the Pipe’s complimentary storage (if enabled) and to your own storage (if configured), together with all the other files related to a recording.

The transcription JSON file has the following structure:

{
    "schema_version": 1,
    "recording_id": 123,
    "recording_name": "STREAM_NAME",
    "provider": "replicate",
    "model": "vaibhavs10/incredibly-fast-whisper",
    "version": "model_version_hash",
    "prediction_id": "prediction_id",
    "completed_at": "2026-09-02T08:22:00Z",
    "output": {
        "chunks": [
          {
            "text": " The little tales they tell are false the door was barred locked and bolted as well ripe pears are fit hours fly by much too soon. The room was crowded",
            "timestamp": [
              0,
              29.72
            ]
          },
          {
            "text": " with a mild wab. The room was crowded with a wild mob. This strong arm shall shield your",
            "timestamp": [
              29.72,
              38.98
            ]
          },
          {
            "text": " honour. She blushed when he gave her a white orchid The beetle droned in the hot June sun",
            "timestamp": [
              38.98,
              48.52
            ]
          }
        ],
        "text": " The little tales they tell are false the door was barred locked and bolted as well ripe pears are fit hours fly by much too soon. The room was crowded with a mild wab. The room was crowded with a wild mob. This strong arm shall shield your honour. She blushed when he gave her a white orchid The beetle droned in the hot June sun"
    }
}