Veo3 + Tigris Text‑to‑Video (Next.js, TypeScript)

Build a text‑to‑video app using Google's Veo3 model and Tigris object storage. Enter an idea, optionally enhance it into structured cinematic parameters, generate a video with Veo3, and automatically store and serve it from Tigris via presigned URLs.

Example idea:

Cinematic. An Extreme Close-Up of a scientist’s trembling hands holding a test tube filled with glowing green liquid. Sirens blare in the background as red warning lights flash across the lab. She whispers into a recorder, “If this works, humanity has one last chance.”

Requirements

• Node v20.11.0 or later
• A Google AI Studio API key for Veo3 and Gemini 2.5 Flash
• A Tigris (S3‑compatible) bucket and access keys

Project structure

veo3/
  └─ typescript/
     ├─ public/
     ├─ src/
     │  ├─ app/
     │  │  ├─ api/
     │  │  │  ├─ enhance/
     │  │  │  │  └─ route.ts
     │  │  │  └─ generate/
     │  │  │     └─ route.ts
     │  │  ├─ favicon.ico
     │  │  ├─ globals.css
     │  │  ├─ layout.tsx
     │  │  └─ page.tsx
     ├─ eslint.config.mjs
     ├─ next.config.ts
     ├─ package.json
     ├─ postcss.config.mjs
     └─ tsconfig.json

Setup

1. Create a Google AI Studio API key

• Create or select a Google Cloud project
• In AI Studio, click “Get API key” → “Create API key” and link it to your project

2. Create a Tigris bucket and access keys

• Go to https://storage.new and create a bucket (recommended name: veo3)
• Create Read/Write access keys and copy the credentials
• Ensure your user/role has Editor permissions on the bucket

3. Configure environment variables

Create a .env.local in veo3/typescript with:

GOOGLE_API_KEY=your_google_ai_studio_key
TIGRIS_STORAGE_ACCESS_KEY_ID=your_tigris_access_key_id
TIGRIS_STORAGE_SECRET_ACCESS_KEY=your_tigris_secret_access_key

Notes:

• The API runs server‑side only; do not expose keys client‑side.
• The S3/Tigris configuration in api/generate/route.ts uses:
  • Bucket name: veo3 (change S3_BUCKET_NAME in code if you prefer another)
  • Region: auto
  • Endpoint: https://t3.storage.dev
  • Path style: false
  • Presigned URL expiry: 1 hour

Install and run

npm install
npm run dev
# open http://localhost:3000

How it works

• UI (src/app/page.tsx)

  • Enter a raw idea and click “Enhance” to extract cinematic parameters using Gemini 2.5 Flash
  • Optionally tweak parameters (subject, action, scene, camera angle/movement, lens effects, style, temporal elements, sound effects, dialogue)
  • Click “Generate Video” to send a prompt to Veo3 and preview the result
  • Shortcuts: Cmd+E (Enhance), Cmd+K (Toggle Parameters), Cmd+Enter (Generate)

• Enhance API (src/app/api/enhance/route.ts)

  • POST body: { "prompt": string }
  • Uses gemini-2.5-flash to return structured parameters
  • Response: { ok: true, structured, params }

• Generate API (src/app/api/generate/route.ts)

  • POST body: { prompt: string, numberOfVideos?: number }
  • Calls Veo3 (veo-3.0-generate-preview) via @google/genai, polls until done
  • Downloads generated video bytes, uploads to Tigris (S3 API), returns presigned URL(s)
  • Response: { url: string[] } (UI uses the first item)

Typical generation takes ~1 minute. Inference has a cost; see Google AI Studio pricing.

API examples

Enhance prompt:

curl -sS -X POST http://localhost:3000/api/enhance \
  -H 'Content-Type: application/json' \
  -d '{"prompt":"An astronaut planting a flag on a neon-lit asteroid."}'

Generate video:

curl -sS -X POST http://localhost:3000/api/generate \
  -H 'Content-Type: application/json' \
  -d '{"prompt":"Cinematic, astronaut on neon-lit asteroid, close-up, dolly in, lens flare.", "numberOfVideos":1}'

Tigris configuration

This app talks to Tigris using the AWS SDK for S3. By default:

• Bucket: veo3
• Endpoint: https://t3.storage.dev
• Region: auto
• forcePathStyle: false
• Presigned URL expiry: 3600 seconds

To change the bucket name or expiry, edit S3_BUCKET_NAME and the expiresIn value in src/app/api/generate/route.ts.

Troubleshooting

• 400 Missing prompt: Provide a non‑empty prompt in the POST body
• 500 S3 is not configured: Ensure Tigris env vars are set and the bucket exists (default veo3)
• 403 from Tigris: Verify access key permissions and bucket rules (Editor)
• Fetch video failed: Ensure GOOGLE_API_KEY is valid and Veo3 access is enabled for your project
• Empty/expired playback: Presigned URLs expire after 1 hour; regenerate if needed

Acknowledgements

• Google AI Studio (Veo3 and Gemini 2.5 Flash)
• Tigris S3‑compatible storage (https://storage.new)