This is an n8n workflow that turns a topic or transcript into a finished video without manual editing. It combines DeepSeek AI scriptwriting, text-to-speech voiceovers, AI-generated visuals and automatic video assembly.
The workflow starts with a form where the user enters a topic, the desired duration, a TTS voice and a visual style such as Pixar, Lego or Cyberpunk. DeepSeek AI, accessed through OpenRouter, writes a script with a title, description, hook and call-to-action. The script is converted to audio with the chosen TTS voice, and timestamped captions are generated for synchronization.
The script is then split into scenes of 5 to 6 seconds. Each scene becomes an image prompt in the selected style and is sent to together.ai or fal.ai to generate the image. The images become video clips with Ken Burns-style zoom effects timed to the script. The clips are joined into one video with captions overlaid. The finished video uploads to Google Drive and Telegram, and progress is tracked in Google Sheets.
- Platform
- n8n workflow automation
- Scriptwriting
- DeepSeek AI via OpenRouter
- Voiceover
- Text-to-speech with customizable voices
- Image generation
- together.ai or fal.ai
- Input
- Form with topic, desired duration, TTS voice and visual style
- Visual styles
- Pixar, Lego, Cyberpunk and others
- Scene length
- 5-6 seconds per scene
- Video effects
- Ken Burns-style zoom, overlaid captions
- Output
- Uploaded to Google Drive and Telegram
- Progress tracking
- Google Sheets
- n8n nodes
- Set, Code, Wait, Switch, Split Out, Telegram
- Integrations
- OpenRouter, together.ai, fal.ai, Google Drive, Google Sheets, Telegram