👋 Welcome to the SparkAi System!
🚀 The new generation AIGC one-stop B/C end solution! AI large model questioning, AI painting, AI video generation (latest release), document analysis, multi-modal image recognition and understanding, TTS & voice recognition dialogue, AI face swapping, support for AI agent applications (supports GPTs applications + Prompt custom preset applications), plugin system, AI music generation, a one-stop system!! Supports all OpenAI models, Midjourney full-function painting, Claude3, iFlytek Spark, Doubao (Doubao-lite-4k) large models, Stable Diffusion, DALL·E (DALL-E2/E3/E4 painting), ChatGLM, Tongyi Qianwen, Tencent Hunyuan, 360 ZhiNao, Baichuan AI, Volcengine, New Bing, Gemini, Moonshot, etc. Supports dialogue session isolation, cloud storage, and data sharing across multiple devices! Supports flexible billing and various plan modes, and other core functions, complete AI integration + user backend management and data statistics.
✨ Commercial Version Private Deployment (license authorization)
The private deployment version is mainly for webmasters and companies intending to build an AI platform. If you just want to use the official SparkAi site, you can skip this.
🎉 System Features (All the following functions have been developed and are supported!)
- 🚀 Uses Node.js + NestJs server-side rendering technology, supports high concurrency (tens of thousands or even tens of millions of users requesting at the same time); provides higher availability and stability.
- 📡 Supports regular Baota deployment and Docker one-click deployment. All integration configurations can be completed on the backend interface.
- 🤖 Supports OpenAI official API + One API and other relay integrations (i.e., supports all OpenAI models, third-party large models, and all domestic models); supports OpenAI / Azure OpenAI / Anthropic Claude / Google Gemini / Midjourney / iFlytek Spark / Baichuan AI / Moonshot / Zhipu ChatGLM / Tongyi Qianwen / Lark large models / Tencent Hunyuan / 360 ZhiNao and other models, adapts to LocalAI / Ollama and other models (LLaMa / RWKV / SDXL), supports integration with the latest gpt-4o large model released by OpenAI.....
- 🖥️ Custom model integration system: All models follow OpenAI's input format. When OpenAI releases a new language model, it can be added in the backend immediately without requiring system development for each integration! Similarly, it supports relay systems that have integrated other third-party models, meaning that the latest models released by OpenAI, third-party large models, and domestic large models can be quickly made available. Dialogue models can be custom classified, named, sorted, and have some functions customized. 5.🎞️ Independent AI video generation function, supporting text-to-video/image-to-video (newly added).
- ✨ Supports any OpenAI format relay API, self-developed channel load balancing and distribution algorithms, supports multi-channel high-concurrency call management, and supports multiple API Key polling! (Priority/weight/status management).
- 🤯 Dialogue session isolation, cloud storage, and data sharing across multiple devices!
- 🎈 Multi-modal model support: Image recognition and document analysis are supported. Supports custom dialogue models to enable multi-modal support.
- 🌏 Supports networking mode, allowing the model to be extended to search and summarize real-time content from the current network.
- 🎃 Supports AI agent applications: GPTs applications + Prompt custom preset applications; GPTS: supports custom addition of GPTS in the backend, and can also be searched site-wide = official search.
- 👂 Voice recognition & TTS, supports OpenAI / Azure and other voice recognition & TTS models, supports Whisper & TTS format relay; supports TTS dialogue input and reply mode.
- 🎨 Supports Midjourney full functionality, including Imagine / Upscale / Vary / Zoom Out and many other functions, and supports Seed acquisition. Supports relay API / partial redrawing (Vary Region), etc.
- 💭 Professional painting supports text-to-image, image-to-image, reference image mixing, character consistent reference images, style consistent reference image generation, etc., supports AI face swapping, and supports image mixing.
- 🪄 Dall-E-2/3 painting support, supports all parameters and has a separate module.
- 🔥 Powerful Markdown syntax support (supports code highlighting / LaTeX formulas / Mermaid mind maps / chart drawing), supports integrated dialogue drawing models (DALL-E / Stable Diffusion / Midjourney, etc.).
- 📖 Out-of-the-box document parsing service, supports parsing of file types such as Pdf / Docx / Xlsx / audio / images.
- 🗂 Storage system: Supports local, Alibaba Cloud OSS, Tencent Cloud COS, and Chevereto image hosting.
- 💰 Membership system: Provides three currency types: regular model points, advanced model points, and painting points. Supports custom deduction modes and quotas for various products. Supports custom deduction fee types for models, and multiple billing methods: time-limited, unlimited, and custom combination package settings (free / point (times) deduction billing).
- 💸 Supports official WeChat payment, EasyPay, CodePay, HupiJiao Pay and other payment methods, supports synchronous checking of order status, and supports order search and management.
- 👋 Mall system: Custom product packages, can be customized to generate permanent packages and limited-time packages.
- 🔶 Check-in system: Supports configuring daily check-ins to give different amounts of rewards.
- 📦 Visitor system: Supports using some content without logging in, and can experience the site's functions in guest mode.
- 🌏 Distribution system: Supports A + B distribution mode, the backend can customize the distribution commission amount, and can be set separately for individual users. Supports setting withdrawal thresholds, and supports multiple withdrawal methods for users (Alipay, WeChat, bank card, etc.).
- 🎫 Card and password system: Supports batch generation of cards and passwords for users to redeem. The user end can redeem cards and passwords or purchase from third-party card networks. Supports a complete card and password redemption code system, supports custom points, supports batch generation and card and password redemption code management, supports viewing the user who claimed the card and password redemption code / creation time / claim time and other information; deleting card and password redemption codes and other functions.
- 🎉 Login system: Supports different login and registration methods such as email | mobile SMS | WeChat QR code.
- 🏅 The plugin system is open, with various built-in plugins and more plugins will be continuously developed.
- 🖥 Multi-terminal support, adaptive layout, adapts to PC, mobile, tablet and other devices. Supports Web / PWA / App / Mini Program (to be developed later), UI mobile adaptation, supports switching between light and dark themes, supports Windows / MacOS / Linux / Android / iOS App.
- ✔️ Professional site management system, supports custom configuration of website LOGO / site name / footer / Baidu statistics / AI name / copyright information / contact information / site announcement / welcome message and other content; supports setting user initial points, separate user group configuration, etc.
- 📈 Data statistics, rich and beautiful dashboard statistics, supports daily new user count statistics / total and daily dialogue count / total and daily new painting count / total and daily order count statistics / and other information statistical chart trend display.
- 📄 Multiple restrictions on sensitive content, supports custom sensitive words / Baidu content review and other security checks, which can be enabled together. Supports sensitive word check record query and statistics!
- 🧩 Dynamic menu system: Supports embedded web pages, external link jumps, internal path jumps, can customize switch menus, add menus, custom ICONs, custom menu types, can use third-party site embedding, and supports separate settings for mobile and PC.
- 📡 The management end supports super administrator and regular demo accounts for the distribution system, with regular accounts having read-only permissions.
- ✈ The management end supports configuring keyword replies, and can customize keywords and reply content.
- 📣 Promotion invitation rewards, supports setting custom point rewards for inviting and invited users.
- 📝 Section addition, supports adding custom sections embedded in the website page, supports setting section icons, making the site richer and more diverse.
- 🚥 Other functions
- 🏅 More AI capabilities are continuously being developed and updated...
System Technical Architecture
- Front-end: Vite + vue3 + ts + NaiveUI + Tailwindcss
- Management end: Vite4 + Vue3 + Element-Plus
- Server-side (backend): Nodejs + NestJs
- Data support: Mysql5.7(+) + Redis
- Operating environment: Linux, Windows, MacOs (Linux is recommended)
- Data storage: Local storage | Alibaba Cloud OSS object storage | Tencent Cloud COS object storage | Chevereto image hosting
Server Requirements
The minimum server configuration requirement is 1C1G, and 2C4G or above is recommended for high concurrency.
Operating Environment
- Docker
- Linux
- Windows
- MacOS Server
- Kubernetes
- Supports ARM64 & X86 (32/64) architecture
🎉 系统功能特性（以下功能均已开发支持！）
- 🚀 使用 Nodejs + NestJs 服务端渲染技术, 支持高并发（万级甚至千万级用户同时请求使用）；提供更高的可用性和稳定性。
- 📡 支持宝塔常规部署和Decker一键部署方式，所有对接配置均可在后台界面上完成。
- 🤖 支持OpenAI官方API + One API 等中转对接使用（即支持OpenAI全模型、三方大模型、国内全模型）；支持 OpenAI / Azure OpenAI / Anthropic Claude / Google Gemini / Midjourney / 讯飞星火 / 百川 AI / Moonshot / 智谱 ChatGLM / 通义千问 / 云雀大模型 / 腾讯混元 / 360 智脑等模型, 适配 LocalAI / Ollama 等模型 (LLaMa / RWKV / SDXL )，支持openai发布最新gpt-4o大模型对接使用.....
- 🖥️ 自定义模型对接系统：所有模型走OpenAI的输入格式，当OpenAI发布新的语言模型，可后台第一时间添加，无须系统每次开发对接才能使用！同理支持对接了其他三方模型的中转系统，即支持OpenAI全模型、三方大模型、国内大模型 最新发布模型快速可用。对话模型自定义分类、自定义名称自定义排序、以及部分功能自定义。 5.🎞️ AI视频生成独立功能，支持文生视频/图生视频（最新新增）
- ✨ 支持任何 OpenAI 格式中转API, 自研渠道均衡负载和分配算法, 支持多渠道高并发调用管理， 支持多 API Key 轮询！ (优先级/权重/状态管理)。
- 🤯 对话会话隔离、云端存储、支持多设备共享数据！
- 🎈 多模态模型支持： 识图、文档分析支持。支持自定义对话模型开启多模态支持。
- 🌏 支持联网模式、可对模型进行扩展搜索当前网络实时内容总结。
- 🎃 支持AI智能体应用：GPTs应用+Prompt自定义预设应用；GPTS：支持GPTS后台自定义添加、也可以全站搜索 = 官方搜索。
- 👂 语音识别 & TTS , 支持 OpenAI / Azure 等语音识别 & TTS 模型，支持 Whisper & TTS 格式中转；支持TTS对话输入回复模式。
- 🎨 支持 Midjourney全功能, 支持 Imagine / Upscale / Vary / Zoom Out 等众多功能， 支持 Seed 获取。支持 中转API / 支持局部重绘 (Vary Region)等等。
- 💭 专业绘画支持文生图、图生图、垫图混图、角色一致参考图、风格一致参考图生成等、支持AI换脸、支持混合图片。
- 🪄 Dall-E-2/3绘画支持、支持所有参数并享有单独模块。
- 🔥 强大 Markdown 语法支持 (支持 代码高亮 / LaTeX 公式 / Mermaid 思维导图 / 图表绘制), 支持集成对话绘图模型 (DALL-E / Stable Diffusion / Midjourney 等)。
- 📖 开箱即用的文档解析服务, 支持 Pdf / Docx / Xlsx / 音频 / 图片等文件类型解析。
- 🗂 存储系统：支持本机、阿里云OSS、腾讯云cos、Chevereto图床多种类型。
- 💰 会员系统： 提供普通模型积分、高级普通模型积分、绘画积分三种货币类型、支持各类商品自定义扣费模式与额度。支持自定义模型扣除费用类型、多种计费方式：按时间限制、无时间限制、自定义组合套餐设置 (免费 / 积分（次数）扣费 计费)。
- 💸 支持官方微信支持、易支付、码支付、虎皮椒支付等支付方式, 支持同步检查订单状态, 支持订单搜索和管理。
- 👋 商城系统：自定义商品套餐、可自定义生成永久套餐、限时套餐。
- 🔶 签到系统： 支持配置每日签到赠送不同额度的奖励。
- 📦 访客系统： 支持不登录使用部分内容、可游客模式体验站内功能。
- 🌏 分销系统：支持 A + B 分销模式、后台可自定义分销提成额度、可对单独用户单独设置。支持设置提现门槛选项，支持用户多种提现方式选择（支付宝、微信、银行卡等）。
- 🎫 卡密系统：支持批量生产卡密供用户兑换、用户端可兑换卡密或三方卡网购买支持完备卡密兑换码体系, 支持自定义点数, 支持批量生成和卡密兑换码管理, 支持卡密兑换码查看领取用户 / 创建时间 / 领取时间等信息；删除卡密兑换码等功能。
- 🎉 登录系统：支持 邮箱 | 手机短信 | 微信扫码 等不同登录注册方式供选择。
- 🏅 插件系统开放、内部支持各类插件并会持续开发更多插件。
- 🖥 多端支持，自适应布局、适配PC、移动端、平板等设备。支持 Web / PWA / App / 小程序（后续开发）, UI 移动端适配, 支持明暗主题切换, 支持 Windows / MacOS / Linux / Android / iOS App。
- ✔️ 专业站点管理系统, 支持自定义配置网站 LOGO / 站点名称 / 页脚 / 百度统计/ AI名称 / 版权信息/ 联系方式 / 站点公告 / 提示欢迎语等多种内容等多种内容;支持设置用户初始点数, 单独用户群配置 等等。
- 📈 数据统计, 丰富美观的仪表盘统计，支持每日新增用户数量统计 / 总计和每日对话数量 / 总计和每日新加绘画数量 / 总计和每日日订单数量统计 / 等多种信息统计图表趋势展示。
- 📄 内容敏感词多种限制、支持自定义敏感词/百度内容审核等安全检测、可共同开启。支持敏感词检查记录查询统计！
- 🧩 动态菜单系统：支持内嵌网页、外部链接跳转、内部路径跳转、可以自定义开关菜单、添加菜单、自定义ICON、自定义菜单类型、可使用三方站点内嵌、支持移动端PC端各自单独设置。
- 📡 管理端支持超管和普通演示账户用于分销系统、普通账号仅可读权限
- ✈ 管理端支持配置关键词回复、可自定义关键词及回复内容
- 📣 推广邀请奖励，支持邀请和被邀请用户设置自定义积分奖励。
- 📝 板块添加, 支持添加自定义板块嵌入网站页面, 支持设置板块图标，使站点更加丰富多样。
- 🚥其他功能
- 🏅 更多AI能力持续开发更新中...
系统技术架构
- 前端： Vite + vue3 + ts + NaiveUI + Tailwindcss
- 管理端：Vite4 + Vue3 + Element-Plus
- 服务端（后端）：Nodejs + NestJs
- 数据支持： Mysql5.7(+) + Redis
- 运行环境：Linux、Windows、MacOs（推荐使用Linux）
- 数据存储：本地存储 | 对象存储阿里云oss | 对象存储腾讯云cos | chevereto图床
服务器要求
最低服务器配置运行要求 1C1G, 高并发推荐使用 2C4G 及以上。
运行环境
- Docker
- Linux
- Windows
- MacOS Server
- Kubernetes
- 支持 ARM64 & X86 (32/64) 架构