GPT Image Playground
A web-based image generation and editing tool built on the OpenAI gpt-image-2.5 API, supporting reference images, mask editing, batch generation and a multi-turn agent chat mode, with data stored locally.
GPT Image Playground is a web application for generating and editing images with OpenAI's gpt-image-2.5 API, or compatible providers such as OpenAI-compatible endpoints, sub2api and fal.ai, plus custom importable HTTP providers. It supports uploading up to sixteen reference images, a visual mask editor, single-run batch generation, and converting a satisfying result back into a reference image to continue iterating.
Its agent mode uses the Responses API to hold a multi-turn conversation with context awareness, letting the agent call an image-generation tool, generate several related images concurrently in one turn, and continue a turn automatically to handle dependent images; edited or regenerated messages create switchable branches, and an optional web-search tool can attach source links. History, favorites and settings are stored only locally in the browser, with waterfall and gallery views, favorite folders and batch ZIP downloads.
It is built with React, TypeScript, Vite and Tailwind CSS, and is licensed under MIT.