AdGen: AI Creative Studio for Brands
Overview
AdGen is an AI creative studio for brands that want to go from a campaign idea to polished ad output in one place. Instead of juggling separate tools for copy, images, edits, and video, teams build campaign-ready creatives from brand context, refine them inside a live canvas, and export platform-specific assets — all from the same workflow. Create flows capture the inputs a brand actually cares about (campaign direction, target platform, logos and product images), so the output feels tailored for YouTube, Reels, or TikTok rather than generic.
Multi-agent architecture
AdGen is stitched together as a coordinated creator pipeline. Gemini is the reasoning layer that carries campaign context from one specialist agent to the next; still-image generation and editing run on Nano Banana Pro, and final video rendering on Veo.
- Brand Analyzer — reads logo and product images and extracts the brand's visual identity.
- Creative Director — turns the brand analysis and brief into a master creative brief that every downstream agent follows.
- Copywriter — generates platform-aware messaging in sync with the established direction.
- Art Director — translates the brief into visual instructions for composition, style, lighting, and product placement.
- Image generation & editing — Nano Banana Pro produces the still creative and applies targeted canvas edits without discarding the composition.
- Video Director & generation — converts the same brief and stills into motion direction, rendered by Veo for the selected platform.
Tech stack
| Frontend | Next.js 16, React 19, TypeScript, MUI, Konva, Zustand |
| Backend | Node.js, Express, TypeScript |
| AI orchestration | Google GenAI SDK — Gemini, Nano Banana Pro, Veo |
| Platform | Firebase Auth, Cloud Firestore, Cloud Storage, Cloud Run |
| Infra & CI/CD | Terraform, GitHub Actions, Workload Identity Federation |