One frontend for ChatGPT, Midjourney, Suno, and a dozen other AIs

What it does This project is a JavaScript frontend that bundles access to a sprawling list of generative AI services—ChatGPT, Midjourney, GPTs, Suno, Udio, Runway, Luma, Pika, Flux, Ideogram, Kling, and others—into a single interface. It began as a fork of ChenZhaoYu’s chatgpt-web and grew into a multi-modal control panel. The codebase runs as a web app, PWA, or standalone desktop client on Linux, Windows, and macOS, and is deployable through Vercel or Docker. The README is candid about its architecture: it is essentially an aggregation layer built atop midjourney-proxy, Suno-API, and Luma-API.

The interesting bit The breadth is almost comical. Instead of treating each generative model as a separate tool, the UI folds them into one utility grid, with menu tabs for music, video, dance, and real-time voice chat alongside standard text and image generation. It even exposes niche Midjourney operations like pan-and-zoom, face swapping via InsightFace, and seed extraction—features usually trapped inside Discord bots. The project openly recommends using a third-party API aggregator so that a single key and endpoint can power the entire suite.

Key highlights

• Supports text (ChatGPT/GPTs), image (Midjourney, Flux, DALL·E, Ideogram), video (Runway, Luma, Pika, Kling), music (Suno, Udio), and dance generation (Viggle) in one interface
• Includes OpenAI Realtime voice conversations, browser-native speech recognition, and TTS/Whisper integration
• Midjourney features cover inpainting, upscaling, directional panning, face replacement, image blending, and seed retrieval
• Runs as a web app, PWA, or desktop client; deployable through Vercel or Docker
• Desktop builds and Docker deployments support file uploads and brute-force protection, while Vercel deployments do not

Caveats

• It is explicitly a secondary fork and glue layer: you will need separate backend proxies (midjourney-proxy, Suno-API, Luma-API, etc.) running for the corresponding features to work
• Several capabilities—such as Cloudflare R2 file storage and anti-brute-force authentication—are marked as unsupported on Vercel and require Docker or desktop installs
• The “one key” convenience relies on a specific recommended commercial aggregator (openai-hk.com), introducing a hard dependency if you want unified billing

Verdict Consider it if you are self-hosting multiple generative APIs and want a single dashboard instead of a browser tab farm. Skip it if you need a lightweight, single-purpose ChatGPT clone or prefer native first-party interfaces.