AI Video & Podcast Transcription Workflow
Transcribe long-form audio or video, clean the verbatim text, and produce both a readable transcript and a timed subtitle track.
No credit card for free-tier templates · every node inspectable · export or fork at any time
中文What this workflow does
Upload video/audio → 5-node: denoise → transcribe → speaker ID → key insight extraction → multi-format output. Rivals Descript $24/mo.
A polished transcript document and a synced subtitle file.
Pipeline breakdown
5 nodes, executed in order. Every step is white-box — inspect the model, the prompt and the intermediate payload.
- 01
Pillow
Image pre-processing — deskew, denoise, crop and normalise before analysis
audio→audio - 02
Faster-Whisper
Local speech-to-text transcription with timestamps and speaker turns
audio→json - 03
LLM Reasoning
Large-language-model step that extracts, classifies, validates or writes structured output
json→json - 04
LLM Reasoning
Large-language-model step that extracts, classifies, validates or writes structured output
json→json - 05
Python-DOCX
Renders the final Word document with headings, tables and styling
json→docx
Built for
视频内容多平台分发
播客逐字稿生成
采访录音整理
在线课程字幕
FlowSync vs Descript / Rev vs generic workflow builders
| FlowSync template | Descript / Rev | n8n / Zapier / Make | |
|---|---|---|---|
| Time to first result | Minutes — pipeline pre-built | Minutes, single-purpose | Hours — you design the flow |
| Intelligence included | Prompts, validation, output schema tuned | Fixed product logic | None — you write it |
| Editable pipeline | Every node exposed | Closed product | Fully editable |
| Chainable with other AI skills | 5 skills here, 50+ available | Limited | Via API glue only |
| Output format | Business deliverable (Excel / DOCX / MP4) | Product-native export | JSON, then DIY |
The short version: workflow builders give you a canvas, point products give you one fixed answer. A FlowSync template gives you a working pipeline you can still open up and change.
Field notes
Descript $24/月，我一个月只出 4 期视频。这个免费版效果一样，金句提取功能 Descript 还要加钱。
采访录音自动转文字+标时间戳，写稿效率提升太多了。
FAQ
Is the Video/podcast → transcript template free to run?
Yes. This template is on the free tier — create an account and run it without a card. Usage limits apply to concurrent runs, not to the template itself.
How is this different from building the same flow in n8n or Zapier?
5 pre-configured nodes (Pillow → Faster-Whisper → LLM Reasoning → LLM Reasoning → Python-DOCX) with prompts, validation and output formatting already tuned for the task. You import it and get the deliverable, not a blank canvas.
Can I edit the pipeline or swap a node?
Yes. Every node is exposed in the editor — change the model, rewrite a prompt, insert or remove a step, then save it as your own private template. Nothing is a black box.
What do I actually get back?
A polished transcript document and a synced subtitle file.
What happens to the files I upload?
上传的音视频仅在转录过程中临时使用，完成后即刻清除。文字稿仅存储于您的账户。语音数据不会被用于任何模型训练或第三方共享。
Data handling
上传的音视频仅在转录过程中临时使用，完成后即刻清除。文字稿仅存储于您的账户。语音数据不会被用于任何模型训练或第三方共享。
中文说明 · 视频/播客 → 文字稿（对标 Descript）
视频/播客 → 文字稿（对标 Descript）（音视频 · 免费）
上传视频/音频 → 5 节点：降噪 → 转写 → 说话人识别 → 要点提炼 → 多格式输出。对标 Descript $24/月。
管线共 5 个节点：Pillow → Faster-Whisper → LLM Reasoning → LLM Reasoning → Python-DOCX。每个节点均可在编辑器中查看与修改，支持另存为你自己的私有模板。
Related templates
AI Video Generation Workflow Template
Turn any text, file or URL into a finished narrated video — storyboard, voiceover script, subtitles, 16:9 + 9:…
Open → FinanceAI Invoice Processing Agent Template
Drop in invoices or statements in any format — OCR, field extraction, amount validation, compliance check and …
Open → OfficeAI Meeting Notes Agent Workflow
Recording in, decisions out. Transcribe, separate speakers, extract decisions and action items, then render a …
Open →Run it, then make it yours
Import the pipeline, swap a model, rewrite a prompt, save it private. FlowSync is the execution layer — the template is just the starting point.