mylxsw/aidea-server

AIdea 是一款支持 GPT 以及国产大语言模型通义千问、文心一言等，支持 Stable Diffusion 文生图、图生图、 SDXL1.0、超分辨率、图片上色的全能型 APP。

AIdea Server – What It Is

AIdea Server is an open‑source backend written in Go that powers the AIdea client app. It provides a self‑hostable API for chat‑style large language models (LLMs) and image‑generation models, exposing an OpenAI‑compatible REST interface so existing tools that speak the OpenAI protocol can call it directly.

Core Features

Chat & Completion API – Wraps multiple LLM providers behind a unified OpenAI‑compatible endpoint, supporting streaming responses.
Image Generation – Integrates popular image‑generation models (e.g., Stable Diffusion‑style services) through the same API surface.
User & Billing Management – Built‑in token accounting, pricing tables, and payment integrations (Alipay, Apple, etc.) to monetize usage.
Multi‑Channel Notifications – DingTalk bot, SMS (Alibaba Cloud, Tencent), and email alerts.
Auxiliary AI Services – Text‑to‑speech (Qiniu), speech‑to‑text (Tencent), translation (Youdao).
Task Queue & Scheduler – Asynchronous jobs (e.g., daily usage stats) run via an internal queue system.
Rate Limiting & Security – JWT authentication, per‑user rate limits, and content‑safety checks via Alibaba Cloud services.
OpenAI‑compatible API – Any third‑party client that understands the OpenAI API can talk to AIdea Server without modification.

Architecture & Code Layout

Directory	Purpose
`api`	Implements the OpenAI‑compatible endpoints used by external tools.
`server`	Endpoints specific to the AIdea client application.
`cmd`	Main entry point (`main.go`).
`config`	Configuration structs and default `config.yaml`.
`pkg`	Public Go packages that can be imported by other projects.
`ai/`	Provider‑specific model adapters (chat, image, etc.).
`repo/`	Data‑access layer built on the in‑house Eloquent ORM (MySQL, etc.).
`service/`	Business logic that doesn’t belong in controllers or repos.
`queue/`	Simple task‑queue implementation with consumer workers.
`payment/`, `coins/`	Billing, pricing tables, and virtual‑currency handling.
`misc`, `rate`, `token`, `uploader`, …	Utility modules (rate limiting, JWT, file upload, etc.).

The project relies on two internal frameworks:

Glacier Framework – a modular Go framework with dependency‑injection built on go-ioc.
Eloquent ORM – a code‑generation‑based ORM inspired by Laravel, handling DB migrations and models.

Getting Started / Self‑Hosting

Clone the repo and edit config.yaml (or supply environment variables) to point to your database, Redis, and any AI provider keys.

Build:

go build -o build/debug/aidea-server cmd/main.go

Run the binary directly, or use the provided Docker deployment repo (mylxsw/aidea-docker).
Optional – Follow the detailed guide in docs/deploy.md for systemd or Nginx setup, or request assisted deployment via docs/deploy‑vip.md.

A live demo instance is hosted at https://aidea.aicode.cc for quick testing.

Community & Support

WeChat discussion group – QR code in the README; join by adding the ID x-prometheus.
WeChat Official Account – Broadcasts and updates.
GitHub – Issues and pull requests for the server, client, and Docker repo.

License

MIT – free for personal, academic, or commercial use.

TL;DR

AIdea Server is a Go‑based, self‑hostable backend that aggregates LLM and image‑generation services behind an OpenAI‑compatible API, adds user management, billing, and various notification channels, and can be deployed via Docker or directly on a Linux host.