Case study · 2026
Gemini Image Captioner
Scroll
Gemini Image Captioner
View live site- Year
- 2026
- Type
- AI-powered image caption generator
- Tech stack
- Next.js | TypeScript | Ollama | Vercel AI SDK | Tailwind CSS | shadcn/ui | Google Gemini API
- Status
- Featured project
Generate AI-powered captions for your images.
Batch-upload a folder of images, add a custom prefix/suffix, and export a ready-to-train ZIP of captions using Gemini or a local Ollama vision model — built as a drop-in LLaVA replacement for LoRA dataset prep.
Scope & impact
- Dual model support — cloud captioning via the Gemini API or fully local/offline captioning through Ollama
- Batch processes multiple images at once with adjustable detail level and caption prefix/suffix
- Exports all generated captions as a single ZIP, ready to drop into fal or Replicate LoRA trainers