Text to 360°

Generate immersive 360° equirectangular panoramas from a text prompt or a reference image, and explore them directly in the browser with a draggable panoramic viewer.

Powered by fal.ai using OpenAI's gpt-image-2 and gpt-image-2/edit models.

How it works

• Text → 360°: your prompt is wrapped as 360 equirectangular image of <your scene>. Seamless 2:1 aspect ratio, ... and sent to openai/gpt-image-2.
• Image → 360°: drop a reference image and it's sent to openai/gpt-image-2/edit with a panorama-expansion instruction.
• The returned image is rendered onto an inverted sphere in three.js so you can drag to look around and scroll to zoom.

Setup

cp .env.example .env.local
# put your key in .env.local
# FAL_KEY=fal_...

pnpm install   # or npm install / yarn
pnpm dev

Open http://localhost:3000.

Routes

Path	Description
/	Main UI: viewer, prompt bar, sidebar
/api/generate	POST — { prompt, imageUrl?, quality? } → { url }
/api/upload	POST — multipart file upload to fal.storage

Stack

• Next.js 15 (App Router, Node runtime for API routes)
• React 19
• Tailwind CSS 4
• three.js (panorama sphere)
• @fal-ai/client for queue + storage