Infomaniak AitoolsInstall

Summary

Transcribes an audio file using the Infomaniak AI Tools OpenAI-compatible API.

Inputs

• authentication (required) — Choose OAuth2 or API Key authentication.
• path_product_id (required) — Infomaniak AI Tools product identifier (numeric).
• body_file (required) — The audio file to transcribe (base64 string or binary data). Maximum 50 MB, allowed types: mp3, mp4, aac, wav, flac, ogg, opus, wma, m4a.
• body_model (required) — Model name to use for transcription (e.g., whisper‑1).
• bodyParameters.body_language — Language of the input audio (ISO‑639‑1 code).
• bodyParameters.body_prompt — Optional text to guide the model's style or continue a previous segment.
• bodyParameters.body_response_format — Output format: json, text, srt, vtt, or verbose_json (default: json).
• bodyParameters.body_timestamp_granularities — Timestamp granularities to include (word and/or segment). Requires response_format=verbose_json.
• bodyParameters.body_append_punctuations — Punctuation symbols to merge with the previous word (when timestamp_granularities includes word).
• bodyParameters.body_prepend_punctuations — Punctuation symbols to merge with the next word (when timestamp_granularities includes word).
• bodyParameters.body_highlight_words — Underline each spoken word in srt/vtt output (requires timestamp_granularities:word).
• bodyParameters.body_chunk_length — Maximum active segment duration in seconds (useful for subtitle tasks).
• bodyParameters.body_max_line_count — Maximum lines per segment in srt/vtt output (requires timestamp_granularities:word).
• bodyParameters.body_max_line_width — Maximum characters per line before line break in srt/vtt output (requires timestamp_granularities:word).
• bodyParameters.body_max_words_per_line — Maximum words per segment in srt/vtt output (requires timestamp_granularities:word).
• bodyParameters.body_no_speech_threshold — Threshold for treating a segment as silent based on no_speech probability and log probability.
• returnFullResponse — If true, returns the full API response; otherwise returns only the transcription result.

Output shape

array of objects, each with a json field containing the transcription result.

When returnFullResponse is false (default), the node extracts the data property from the API response (or the whole response if no data exists). If the result is an array it is returned as‑is; otherwise it is wrapped in an array. The format of the transcription is determined by body_response_format (default: json).

Examples

Example 1: Simple transcription to JSON

Set authentication, product_id, provide the audio file in body_file, specify model (e.g., whisper‑1). Leave other fields blank.

Example 2: Transcription with subtitles

Same as above, plus set body_response_format to srt or vtt, and optionally enable body_highlight_words, set body_chunk_length to 5, and provide a language code.

Links

• n8n community nodes installation guide