Aliyun Bailian | 阿里云百炼Install

Summary

Perform recorded speech recognition on one or more remote audio files using the Fun-ASR API models.

Inputs

• Model ID | 模型 ID (required) — Specifies the Fun-ASR recorded-speech recognition model ID to use for transcription (e.g., fun-asr, fun-asr-mtl, paraformer-v2). Refer to the Fun-ASR documentation for supported models.
• Audio URLs | 音频地址列表 (required) — A list of HTTP/HTTPS/OSS URLs pointing to audio files to be transcribed, one URL per line. The API accepts only remote URLs, not direct binary uploads.
• Wait for Result | 等待识别完成 — Determines whether to poll the Fun-ASR task until the transcription completes.
• Poll Interval (MS) | 轮询间隔 (毫秒) — If polling, the interval in milliseconds between status checks.
• Max Wait (S) | 最长等待 (秒) — Maximum wait time in seconds before timing out when polling for task completion.
• Endpoint Override | 自定义接口路径 — Optional override for the Fun-ASR REST API endpoint path; defaults to /services/fun-asr-recorded/file-recognition if empty.
• Advanced Options | 高级选项 — Collection of advanced parameters to configure the transcription including channel indices, raw JSON parameters, speaker diarization, speaker count, sensitive word filtering, task ID, and vocabulary ID. These options allow customizing language, punctuation, speaker separation, and accuracy improvements.

Output shape

an object containing the transcription results and metadata from the Fun-ASR API for each audio URL processed

Results include recognized text and may include speaker diarization and timestamp details depending on provided options. When waiting for task completion, the node polls until results are ready or timeout occurs. Errors during processing are reported in the execution flow.

Links

• n8n community nodes installation guide