This ComfyUI tutorial workflow demonstrates end-to-end speech-to-text using the FishAudioSpeechToText node. Audio is loaded with LoadAudio (a sample file, fish_audio_example.mp3, is pre-connected so you can run it immediately), then sent to the Fish Audio ASR service for transcription. The node automatically detects the spoken language across 80+ languages and returns a clean transcript string along with an optional JSON payload of timestamped segments. The transcript is saved to disk via SaveText, while a MarkdownNote provides on-canvas tips.
Technically, LoadAudio decodes your file into an audio tensor and forwards it to FishAudioSpeechToText. When precise_timestamps is enabled on that node, it also returns segments_json containing per-segment text with start/end times (and, when available, finer word-level timing). SaveText writes the main transcript string to your ComfyUI output folder for easy reuse in subtitles, captions, or indexing pipelines. This setup is lightweight, API-driven, and practical for rapid transcription, subtitle prep, or generating text prompts from recorded audio.
API
從程式碼使用此工作流
每個 Comfy 工作流都是一個 JSON 圖。下方的 payload 就是此工作流，與 ComfyUI 執行時完全相同 — 你可以從此 URL 取得、放入版本控制，或在 ComfyUI 載入並逐節點執行。
使用 Comfy SDK 以 TypeScript 或 Python 執行。相同程式碼可用於 Comfy Cloud 或你自架的 ComfyUI — 只需更改 base URL。
// Install (beta)
npm i @comfyorg/sdk
// Run this workflow (TypeScript)
import { Comfy } from "@comfyorg/sdk";
const client = new Comfy({ apiKey: "comfyui-..." });
const wf = await client.workflows.fromFile("workflow_api.json");
const job = await client.run(wf);
await job.getOutputs("<output-node-id>")[0].toFile("output.png");SDK 需使用 API 格式的工作流：請在 ComfyUI 開啟此工作流，並使用 檔案 → 匯出工作流（API）。
