AssemblyAI is a speech AI platform for pre-recorded transcription (99 languages), speaker diarization, PII redaction and Audio Intelligence (sentiment, entities, topics, content moderation). Upload audio or video from any source, transcribe it asynchronously and feed the results into the rest of your flow.
Changelog
1 version
v1.0.0
Latest
- Initial version with UploadFile, TranscribeAudio, GetTranscript, ListTranscripts, FindTranscripts, DeleteTranscript and MakeApiCall components, plus TranscriptCompleted, NewTranscript and TranscriptFailed polling triggers.
Authentication
Triggers
3
New Transcript
Fires when a new transcript is submitted to your account, regardless of its status. Emits a summary of the transcript — use Get Transcript downstream to fetch the full content once it completes.
Transcript Completed
Fires when a transcript finishes processing (status becomes completed). This is the practical way to react to AssemblyAI's asynchronous transcription. Optionally fetches the full transcript for each new completed item.
Transcript Failed
Fires when a transcript ends with an error status. Useful for surfacing failures that would otherwise be invisible in an asynchronous transcription pipeline.
Actions
7
Delete Transcript
Permanently delete a transcript by its id. This removes the transcript text, words and any redacted PII data and is irreversible.
Get Transcript
Retrieve a transcript by its id, including the full text, words, utterances (with speaker labels) and any enabled Audio Intelligence results. If the transcript failed, the component reports the error.
Find Transcripts
Find transcripts filtered by status and/or creation date. Returns a summary of each transcript (the full text is not included, use Get Transcript to fetch content). Returns a maximum of 200 records.
List Transcripts
List transcripts in your account, newest first. Returns a summary of each transcript (id, status, timestamps, audio URL) — the full text is not included, use Get Transcript to fetch content. Returns a maximum of 1000 records.
Transcribe Audio
Submit an audio or video URL for transcription. IMPORTANT: transcription is asynchronous — this returns a transcript id with status "queued", NOT the text. Use the Transcript Completed trigger (or poll with Get Transcript) to retrieve the result once processing finishes.
Upload File
Upload an audio or video file from file storage to AssemblyAI and get back an upload URL to pass to Transcribe Audio. The file is streamed, so large recordings are supported. Note: a file uploaded with one API key can only be transcribed with a key from the same project/region.