The $0 AI Video Workflow — Full Blueprint

You saw the video. Blank screen to finished video in 30 minutes. $0 spent. Here is the exact workflow, every tool, every setting, every prompt.

The 4 Tools (All Free)

| Step | Tool | What It Does |

|------|------|--------------|

| 1 | Nano Banana 2 (higgsfield.ai) | Generates custom images from text prompts |

| 2 | Seedance 2.0 (higgsfield.ai) | Animates still images into video clips (5-15s each) |

| 3 | MiniMax (minimaxi.com) | Generates voiceover from your script |

| 4 | CapCut (capcut.com) | Assembles clips + auto-captions + export |

Step 1: Generate Your Visuals — Nano Banana 2

Go to higgsfield.ai/image/nano_banana_2

• Aspect ratio: 9:16
• Image count: 8
• 2K quality: ON

```

[Subject or scene], [environment], [lighting], [mood], cinematic, ultra detailed, 9:16 vertical, no text, no words, no letters, no watermark

```

```

Crisp hundred dollar bills spread across a dark matte desk, dramatic single overhead light casting sharp shadows, shallow depth of field, cinematic color grading, moody atmosphere, 9:16 vertical, no text, no words, no letters, no watermark

```

```

Smartphone laying flat on a black surface displaying a bright analytics dashboard with green upward graphs, screen glow illuminating surrounding darkness, cinematic, ultra detailed, 9:16 vertical, no text, no words, no letters, no watermark

```

```

Open laptop on a minimalist desk showing lines of code on a dark IDE theme, blue and purple syntax highlighting, soft ambient glow from the screen in a dark room, cinematic, ultra detailed, 9:16 vertical, no text, no words, no letters, no watermark

```

Generate 8 images per prompt. Pick the best 1-2 from each batch.

Step 2: Animate Your Images — Seedance 2.0

Go to higgsfield.ai/seedance/2.0

1. Upload the Nano Banana image you want to animate

2. Write a short motion prompt describing what should move

```

[Describe the motion], [camera movement], [audio direction], cinematic, 9:16 vertical

```

```

The hundred dollar bills shift slightly in the breeze, overhead light flickers subtly, gentle camera drift forward, ambient paper rustling sound, cinematic, 9:16 vertical

```

Each clip generates 5-15 seconds of footage. You need 4-6 clips for a 60-second video.

Step 3: Generate Voiceover — MiniMax

Go to minimaxi.com

1. Write your full 60-second script (aim for 140-160 words)

2. Paste into MiniMax text-to-speech

3. Select voice preset or clone your own voice

4. Generate → download the MP3

• Keep sentences short and punchy
• Write how you talk, not how you write
• Add natural pauses with "..." in the text
• 150 words ≈ 60 seconds of audio

Step 4: Assemble in CapCut

1. Create new project (9:16 aspect ratio)

2. Import all your Seedance video clips

3. Import the MiniMax voiceover

4. Arrange clips on timeline to match the voiceover beats

5. Trim clips to sync with each sentence

6. Enable auto-captions (CapCut does this free)

7. Add any on-screen text overlays

8. Export at 1080x1920, 30fps

The Full Timeline

| Minute | Action |

|--------|--------|

| 0-5 | Write prompts + generate images in Nano Banana |

| 5-15 | Animate best images in Seedance 2.0 |

| 15-20 | Write script + generate voiceover in MiniMax |

| 20-30 | Assemble + caption + export in CapCut |

30 minutes. $0. Done.

Pro Tips

• Batch your prompts. Generate all images first, then animate all at once. Do not switch between tools mid-flow.
• Shoot for 6-8 clips per video. Each clip is 5-10 seconds. Gives you plenty to cut from.
• Keep the voiceover under 60 seconds. Shorter performs better on TikTok/Reels.
• Use the same lighting and mood across all prompts so the video feels cohesive.
• Save your best prompts. You will reuse them across every future video.

What This Replaces

| Old Way | Cost | Time |

|---------|------|------|

| Freelance editor | $100-200 | 2-3 days |

| Voice actor | $50 | 1-2 days |

| Stock footage | $50+ | Hours searching |

| AI workflow | $0 | 30 minutes |

No excuses left. The tools are free. The workflow is here. Go make something.

— Quiv (@big_quiv)