The New Generation AIGC System
🚀{ Progressive } one-stop AIGC system supporting OpenAI's latest flagship model GPT-6 and Anthropic Claude-Opus-5-5 / Claude-Fable-5-1, with free custom integration of the latest models in the backend (no system update needed); 🎨GPT-Image-2.5, 🍌Nano-Banana-2 and Midjourney professional painting, 🎬VEO3.1 video and Seedance2.5 video (coming soon), and 🧠Coze agents (plugins, workflows, functions, knowledge bases); with multi-language support and independent private deployment, it provides comprehensive solutions for individual users (ToC), developers (ToD), and enterprises (ToB).
V6.9.6 refactored the dialogue architecture: automatic routing between the OpenAI Chat Completions / Responses endpoints with built-in fallback retries; the dialogue core decouples upstream and downstream connections, so switching sessions or closing the page does not interrupt generation, and SSE content is saved while generating; supports streaming output, reasoning chain display, web search, a dialogue navigation bar, and Markdown / KaTeX / Mermaid rendering...
Freely add and integrate the latest large models in the admin backend and use them as soon as they are released, with no system update needed; supports the official OpenAI API + any standard chat-format API on the market: OpenAI (GPT-6 / GPT-5.4 / o3 / o4-mini / Codex), Anthropic Claude (Claude-Opus-5-5 / Claude-Fable-5-1), Google Gemini, DeepSeek, Azure OpenAI, plus domestic models such as Doubao, Tongyi Qianwen, Zhipu, Moonshot, iFlytek Spark, and Tencent Hunyuan, and adapts to LocalAI / Ollama...
Standalone GPT-Image-2 / GPT-Image-2.5 painting module (text-to-image, up to 8 reference images, online brush editing, custom quality and resolution); standalone Nano Banana 2 painting module (gemini-3.1-flash-image / gemini-3-pro-image); full Midjourney / Niji features (V7 / V8, Vary Region inpainting, character / style reference images, image blending); DALL·E 2 / 3...
Supports VEO3 / VEO3.1 (with automatic audio) and Sora-2 video generation; Midjourney HD video (high motion / low motion animation); standalone Pika video module (text-to-video / image-to-video); Seedance video via Coze-Agent (first/last frames + sound); standalone Seedance2.5, Kling, and other video modules coming soon...
Integrates Coze agents with plugins, workflows, functions, and knowledge bases, streaming the thinking process and tool calls in real time, with one-click batch import of Coze agents; supports adding GPTs apps from across the web, custom Prompt preset apps, user-created agents, and sharing via link / WeChat QR code / poster...
Supports multi-image understanding (up to 9 images) and photo upload; V6.9.6 added multi-type document understanding for official models (PDF / Word / PPT / Excel / Text / Markdown / HTML), automatically choosing the submission method by file size and token limit; online preview, printing, and download of dialogue files; OpenAI / Azure speech recognition and TTS voice dialogue...
Supports official WeChat Pay (PC Native / JSAPI inside WeChat), EasyPay, CodePay, and HupiJiao Pay; multiple balances (regular / advanced model points, painting points, Agent points) and permanent / limited-time packages; V6.9.6 refactored the points details system, and failed model calls, timeouts, and empty replies can be exempted from charges via a switch, eliminating incorrect and duplicate charges...
Data dashboard (user / dialogue / painting / order trends, model and token usage statistics); model and API key pool management and the large model global configuration center; customizable site LOGO / announcements / ICP number / dynamic menus / carousel, etc.; A + B distribution, invitations, check-in, redemption codes, and guest mode; multiple login methods including email, SMS, and WeChat QR code...
Sensitive words + Baidu content moderation covering dialogue, painting, and mind maps with upfront interception; login device & remote-login detection (IP resolved to city), forced logout on remote login, and forcing specified users offline from the backend; unified sensitive information masking and upload file controls; the System Function Test switch can block all AI generation requests before filing is completed...
Built on a Node.js + NestJS server with Redis caching and an adaptive MySQL connection pool, suited to high-concurrency business scenarios; self-developed channel load balancing and distribution algorithm with multi-channel call management and multi-API-key polling (priority / weight / status management); query optimization for millions of records...
Full system internationalization (frontend UI + backend configuration + backend responses), switching automatically by visitor IP and browser language, with one-click AI internationalization in the admin backend for overseas operation; card style / Notion style themes and AI dialogue immersion mode; adapts to PC / mobile H5 / tablet / WeChat official account, with light/dark themes and PWA...
As of September 2026, the system has been continuously maintained and updated for three and a half years and is still under active development; it includes complete introduction and deployment documentation, supports regular Baota deployment and Docker one-click deployment, and all integration settings can be completed in the admin interface. More AI capabilities are on the way...