1 / 13
AI Code & ProjectsNewMade with Claude Code

GPT 2.5 — AI Image Generator SaaS (Next.js 16 Source Code)

Text-to-image and image-to-image SaaS with credits, Stripe and a gallery. Full source.

AAANKY.AI0.00 reviews0 sales
₹2,499

Commercial license · GST calculated at checkout

Choose a license
  • Secure instant download
  • Free lifetime updates
  • 7-day refund policy
  • Licensed copy · live-site activation, development free

What's included

  • Full Next.js 16 source, TypeScript strict
  • Text-to-image and image-to-image workspace
  • Up to 16 reference images, direct-to-bucket uploads
  • Credits reserved per image, refunded on failure
  • Chats, gallery, favorites, usage and cost
  • Better Auth: passkeys, Google, magic link, codes
  • Plans and prices set in the admin, synced to Stripe
  • Admin: users, plans, blog CMS
  • Licence activation with your purchase code
  • Prisma schema, migrations, demo seed
  • 63 unit tests (Vitest)
  • AGENTS.md, CLAUDE.md, Cursor and Windsurf rules
  • README, FEATURES, CHANGELOG, architecture docs
  • 20 screenshots from the production build
Framework
Next.js 16
Language
TypeScript
Database
PostgreSQL + Prisma
AI
GPT 2.5 Flare, GPT 2.5 Sunburst
Auth
Better Auth
Deploy
Vercel, Any Node host, Docker
File size
18.6 MB
Version
v1.0.1
Updated
Sep 28, 2026

Technical details

Made with

Claude Code

Model: Claude Opus 5.5

Built and documented with Claude Code on Claude Opus 5.5.

Works with

  • Claude Code
  • Codex
  • Cursor
  • Gemini CLI
  • GitHub Copilot
  • Windsurf

AGENTS.md and docs/ai/ let agents change it without scanning the repo.

Try it

Demo accounts
User and admin, one click
Demo generation
Labelled placeholder images, no API cost
Sample workspace
Seeded for each demo account

Licences

Code
Your ANKY.AI tier (LICENSE.md)
Activation
Purchase code on your live site
Dependencies
Their own open-source licences

GPT 2.5 is a complete AI image-generation app: text-to-image and image-to-image with up to 16 reference images, 1K to 4K output, 13 aspect ratios, transparent backgrounds and 1–4 images per click. Every image is a real async task on the Anky AI API, shown as queued and processing until the results fan in, then kept in chats, a searchable gallery and favorites.

Credits are reserved before every call and refunded when an image fails, and plans you price in the admin (synced to Stripe) top them up. Around it sits a production platform: Better Auth with passkeys, Google, magic links and codes; direct-to-bucket uploads verified on the server; per-user rate limits; an admin with users, plans and a blog; and a dark creative workspace with light and mobile layouts.

Run it locally with Docker, two demo accounts and a seeded sample workspace. Demo generation returns labelled placeholder images with no API calls or cost; add an Anky AI API key for real images. AGENTS.md, CLAUDE.md, Cursor and Windsurf rules and a codebase map let Claude Code, Codex or Cursor rebrand and extend it with you.

What you are buying: The full Next.js 16 source of GPT 2.5: the image workspace, the Anky AI API integration, credits and billing, the gallery, the admin, the database schema and demo seed, and the docs. Deploy it on Vercel or any Node host with PostgreSQL.

Third-party costs: Real image generation runs on the Anky AI API and bills your account per image. Stripe, an email provider and S3 or R2 storage are optional and bill you directly. Demo generation costs nothing and needs no key.

GPT 2.5 · AI image generator

Launch your own AI image app
in a weekend, not months

Text-to-image and image-to-image with credits, Stripe plans, a gallery and a dark creative workspace, ready to rebrand and ship.

  • Full source
  • Up to 4K output
  • Credits + Stripe
  • Light, dark & mobile
GPT 2.5 dark creative workspace

13

Aspect ratios

4K

Max resolution

16

Reference images

63

Unit tests

1

One-time payment

Everything included

Everything an image SaaS needs

  • Text-to-image and image-to-image

    Prompts up to 20,000 characters, 1K to 4K output, 13 aspect ratios, transparent backgrounds and 1–4 images per click.

  • Reference images

    Up to 16 references by drag, drop or paste, uploaded straight to your bucket and verified on the server.

  • Credits that refund

    Credits are reserved before every call and returned automatically when an image fails.

  • Chats, gallery, favorites

    Every idea in a chat, every image in a searchable gallery with its prompt, settings and real cost.

  • Safe by design

    Server-only API key, per-user rate limits, at most two active generations and owner checks on every file.

  • Billing and admin

    Plans, prices and credits edited in the admin and synced to Stripe, a customer portal and trial, and an admin for users and a blog.

How it works

From prompt to finished image

  1. 1

    Describe or upload

    Write a prompt or drop in reference images, then pick the model, resolution, ratio and count.

    Image-to-image with reference images
  2. 2

    Watch it work

    Each image is a real async task: queued, processing, then revealed. No fake progress bars.

    Images generating
  3. 3

    Iterate and keep

    Favorite, download or send any result back as a reference. Everything lands in your gallery.

    Gallery of generated images

Who it's for

Ship an image product fast

  • Founders

    Launch a paid image generator with auth, credits, plans and admin already built.

  • Agencies and studios

    White-label an on-brand image tool for clients.

  • Marketing teams

    Campaign visuals, social posts and ad variations in any ratio.

  • E-commerce sellers

    Product shots and lifestyle scenes, with transparent backgrounds for cut-outs.

  • Designers

    Explore concepts quickly, then refine with Use as reference.

  • Communities and schools

    Give members a fair-use image tool with credits and history.

Make it yours

Customize it with Claude Code

The repo ships with AGENTS.md, CLAUDE.md and a codebase map. Open it in Claude Code, Codex or Cursor and ask in plain words:

  • “Rename it Pixelforge, switch the accent to teal and use my logo.”

  • “Charge 2 credits per 1K image and 6 per 4K image.”

  • “Add a Style presets picker that appends a style to the prompt.”

  • “Copy finished images into my own R2 bucket after each generation.”

Project specs

Everything you need to evaluate the stack at a glance.

Framework
Next.js 16
Language
TypeScript
Database
PostgreSQL + Prisma
AI models
GPT 2.5 Flare, GPT 2.5 Sunburst
Auth
Better Auth
Deployment
Vercel, Any Node host, Docker
Version
v1.0.1
Last updated
Sep 28, 2026

Tech stack

Ships with AGENTS.md, CLAUDE.md, GEMINI.md and .cursor/rules.

React 19Tailwind CSS 4shadcn/uiPrisma 7Better AuthStripeS3 / R2 uploadsreact-emailMotionVitest
Package contentsAIinstruction files for coding agents
    • app/dashboard
    • app/api/ai
    • lib/anky
    • lib/image-generation
    • lib/credits.ts
    • components/generator
  • public/demo
  • .anky
  • README.md
  • FEATURES.md
  • CHANGELOG.md
  • LICENSE.md
  • AGENTS.mdAI
  • CLAUDE.mdAI
  • .env.exampleAI
  • docker-compose.yml
  • package.json
Quick start
$ npm install
$ cp .env.example .env
$ docker compose up -d
$ npm run db:migrate
$ npm run db:seed
$ npm run dev

Open the folder in Claude Code, Codex or Cursor and ask it to rebrand, add a feature or deploy — the instruction files explain the architecture and conventions.

#ai image generator saas#text to image app#image to image#nextjs saas source code#ai saas template#credits and stripe billing#image generation dashboard#better auth stripe prisma#shadcn dashboard#claude code ready

More from ANKY.AI

View store