Built by Metorial, the integration platform for agentic AI.
Submit an audio or video file URL for asynchronous transcription with optional audio intelligence features. Supports 100+ languages with automatic language detection, speaker diarization, translation, summarization, sentiment analysis, named entity recognition, chapterization, custom prompts, subtitles, and more. Returns the transcription ID and result URL for polling. Use **Get Transcription** to retrieve results.
Upload an audio or video file to Gladia's servers by providing its URL. Returns a Gladia-hosted URL that can be used with the **Transcribe Audio** tool. Useful for files that require hosting or when working with temporary/authenticated URLs.
Retrieve the post-processed results of a completed live transcription session. Returns the full transcript and any enabled post-processing features like summarization and chapterization.
Retrieve the status and results of a pre-recorded transcription job. Returns the full transcript, utterances with timestamps and speaker labels, and any enabled audio intelligence results (translation, summarization, sentiment, NER, chapters, etc.). Can optionally wait for completion by polling.
Create a new real-time transcription session. Returns a WebSocket URL to stream audio chunks for live speech-to-text. Supports configurable encoding, sample rate, language detection, and real-time audio intelligence features like translation and sentiment analysis.
Permanently delete a pre-recorded transcription and its associated data from Gladia. This action cannot be undone.