# Veo 3
Veo 3 is [[Google]]'s state-of-the-art video generation model — Google's answer to [[Sora]]. It generates high-quality, high-resolution video from text or image prompts, with significantly improved motion realism and audio generation compared to prior versions.
## Key features
- Text-to-video and image-to-video generation
- Native audio generation (sound effects, dialogue, ambient sound) synchronized with video
- High-resolution output (up to 1080p)
- Strong prompt adherence and temporal consistency
- Available via API through third-party platforms
## Access / Platforms
- [[fal.ai]] — API access, fast inference
- [[Replicate.com]] — API access
- Google AI Studio / Vertex AI (direct Google access)
## n8n Workflow: Veo 3 Video Generation
A ready-made [[n8n]] workflow for automated Veo 3 video generation via [[fal.ai]]:
- Source: https://pastebin.com/raw/FyQbNp3X
### Flow
```
Generate prompt
↓
Generate video (fal.ai API call)
↓
Wait 2 min
↓
Check status
├── Complete? No → Wait 2 min → Check status again
└── Complete? Yes → continue
↓
getVideo → downloadVideo
↓
Save to Google Drive
```
### Steps
1. **Generate a prompt** — craft or receive the video prompt
2. **Generate video** — POST to fal.ai Veo 3 endpoint
3. **Wait 2 min** — generation takes time; poll after initial delay
4. **Check status** — loop until job is complete
5. **Download video** — fetch the output URL (`getVideo` → `downloadVideo`)
6. **Save to Google Drive** — store the final video file
## References
- Google DeepMind blog: https://deepmind.google/technologies/veo/
- fal.ai Veo 3 model: https://fal.ai/models/fal-ai/veo3
- Replicate Veo 3 model: https://replicate.com/google/veo-3
## Related
- [[Sora]]
- [[fal.ai]]
- [[Replicate.com]]
- [[n8n]]
- [[Google]]