gemini-audio-transcriber

To transcribe audio to Traditional Chinese with speaker labels, you can use a Node.js script powered by the Google GenAI SDK. It automatically detects multiple speakers and outputs a Markdown transcript suitable for further processing.

For speech-to-text transcription in Node.js, supported audio formats include .mp3, .wav, .ogg, .flac, .m4a, .aac, .webm, and .wma. Inputs can be local files, remote URLs, or data URIs.

Yes, you need a GEMINI_API_KEY to perform speaker diarization and audio transcription. You also need Node.js version 20 or higher installed in your environment to run the script.

Yes, you can transcribe audio directly from a remote URL into a Markdown file. The script accepts local file paths, remote URLs, and data URIs as input to generate the formatted transcript.

The best way to convert an m4a meeting recording into structured notes is to transcribe the audio into a Markdown transcript first, then integrate with a meeting-note-formatter tool to produce the final structured output.

Your Node.js audio transcription script might fail if you are not using Node.js version 20 or higher, or if the GEMINI_API_KEY is missing. Use the dry-run capability to simplify testing and check clear error handling outputs.