GladiaInstall

Summary

Start a transcription job for an audio file accessible via a URL using Gladia's API, with extensive options to configure language detection, translation, custom vocabulary, subtitles, diarization, summarization, and other advanced features.

Inputs

• Audio URL (required) — URL to the audio file to be transcribed.
• Language Config — Configure languages (ISO 639-1) to detect and enable code switching.
• Custom Vocabulary — Whether to use a custom vocabulary for the transcription.
• Custom Vocabulary Config — Details of the custom vocabulary including simple and advanced terms with optional pronunciations, intensity, and language.
• Translation — Enable translation of the transcription output.
• Translation Config — Options for translation including target languages, model type, lipsync, context, and style.
• Custom Spelling — Enable use of a custom spelling dictionary.
• Custom Spelling Config — Custom spelling dictionary entries mapping terms to their variants.
• Subtitles — Enable subtitle generation for the audio.
• Subtitles Config — Subtitle output formats and formatting options such as duration, characters per row, rows per caption, and style.
• Structured Data Extraction — Enable extraction of structured data from the audio.
• Structured Data Extraction Config — Specify data classes for structured data extraction.
• Audio to LLM — Enable sending transcription for language model processing with custom prompts.
• Audio → LLM Config — Prompts to guide the large language model processing of audio transcription.
• Custom Metadata — Arbitrary key-value metadata to include with the transcription job.
• Diarization — Enable speaker diarization to detect distinct speakers.
• Diarization Config — Options to set fixed or range of speaker counts and enhanced diarization mode.
• Summarization — Enable summarization of the transcription output.
• Summarization Config — Type of summarization such as general, bullet points, or concise.
• Callback — Enable callback notification for asynchronous job completion.
• Callback Config — Callback endpoint URL and HTTP method for job status updates.
• Moderation — Enable moderation check on the transcription content.
• Named Entity Recognition — Enable extraction of named entities from the transcription.
• Chapterization — Enable automatic chapter detection in the transcription.
• Name Consistency — Enable name consistency checks across the transcription.
• Sentiment Analysis — Enable sentiment analysis on transcribed content.
• Sentences — Enable segmentation of transcription into sentences.
• Display Mode — Enable enhanced display mode for transcription results.
• Punctuation Enhanced — Enable enhanced punctuation in the transcription.

Output shape

a single JSON object containing the transcription job initiation response from Gladia's API

The output includes details about the transcription job, such as job ID and status. If the callback option is enabled, results will be sent to the specified endpoint asynchronously. Errors produce error messages within the response when continue on fail is enabled.

Examples

Example 1: Start a transcription with translation to multiple target languages

Set 'Audio URL' to the source file, enable 'Translation' with target languages and model, plus optionally enable subtitles or diarization as needed.