Atlassian AI Gateway Proxy

一个Go实现的OpenAI兼容API代理，用于转发请求到Atlassian AI Gateway (Rovo Dev)，具有凭证池管理、轮询重试和指数退避功能。

A Go implementation of an OpenAI-compatible API proxy that forwards requests to the Atlassian AI Gateway (Rovo Dev) with credential pooling, round-robin retries, and exponential back-off.

功能特点 | Features

• OpenAI兼容的API端点 | OpenAI-compatible API endpoints:

  • GET /v1/models – 返回支持的模型列表 | returns supported model list
  • POST /v1/chat/completions – 支持流式和非流式请求 | supports streamed and non-streamed requests
  • GET /health – 健康检查端点 | health check endpoint

• 凭证池管理 | Credential pool management

  • 如果请求失败（401、403或任何5xx错误），会在指数退避后尝试下一个凭证，退避从0.5秒开始，最多到16秒
  • If a request fails with 401, 403 or any 5xx, the next credential is tried after an exponential back-off that starts at 0.5s and doubles up to 16s

• 流式响应支持 | Streaming support

  • 处理流式和非流式聊天完成请求
  • Handles both streaming and non-streaming chat completions

• 错误处理 | Error handling

  • 适当的HTTP状态码和错误消息
  • Proper HTTP status codes and error messages

• Web管理界面 | Web Management Interface

  • 凭证管理（添加、查看、删除）| Credential management (add, view, delete)
  • API令牌管理 | API token management
  • 管理员密码管理 | Admin password management

安装 | Installation

1. 确保已安装Go 1.24.1或更高版本 | Make sure you have Go 1.24.1 or later installed
2. 克隆此仓库 | Clone this repository
3. 安装依赖 | Install dependencies:

     go mod tidy

4. 构建应用程序 | Build the application:

     go build -o atlassian-proxy

配置 | Configuration

首次运行时，应用程序会自动生成一个随机的管理员密码。请在首次登录后立即修改此密码。

When first run, the application will automatically generate a random admin password. Please change this password immediately after first login.

运行 | Running

启动服务器 | Start the server:

  ./atlassian-proxy

或直接使用Go运行 | Or run directly with Go:

  go run .

服务器默认在8000端口启动。您可以通过设置PORT环境变量来更改端口 | The server will start on port 8000 by default. You can change the port by setting the PORT environment variable:

  PORT=3000 ./atlassian-proxy

使用方法 | Usage

服务器运行后，将提供一个OpenAI兼容的API，地址为http://localhost:8000/v1，以及一个Web管理界面，地址为http://localhost:8000/admin/login。

Once running, the server provides an OpenAI-compatible API at http://localhost:8000/v1 and a web management interface at http://localhost:8000/admin/login.

Web管理界面 | Web Management Interface

访问http://localhost:8000/admin/login登录管理界面。首次登录时，使用控制台输出的初始密码。

Visit http://localhost:8000/admin/login to access the management interface. Use the initial password output to the console for first login.

在管理界面中，您可以：

• 管理凭证（添加、查看、删除）
• 生成和查看API令牌
• 修改管理员密码
• 重置管理员密码

In the management interface, you can:

• Manage credentials (add, view, delete)
• Generate and view API tokens
• Change admin password
• Reset admin password

API使用 | API Usage

列出模型 | List Models

  curl http://localhost:8000/v1/models

聊天完成（非流式）| Chat Completion (Non-streaming)

  curl -X POST http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -d '{
    "model": "anthropic:claude-3-5-sonnet-v2@20241022",
    "messages": [
      {"role": "user", "content": "Hello, how are you?"}
    ]
  }'

聊天完成（流式）| Chat Completion (Streaming)

  curl -X POST http://localhost:8000/v1/chat/completions \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_TOKEN" \
  -d '{
    "model": "anthropic:claude-3-5-sonnet-v2@20241022",
    "messages": [
      {"role": "user", "content": "Hello, how are you?"}
    ],
    "stream": true
  }'

支持的模型 | Supported Models

代理支持以下模型 | The proxy supports the following models:

• anthropic:claude-3-5-sonnet-v2@20241022
• anthropic:claude-3-7-sonnet@20250219
• anthropic:claude-sonnet-4@20250514
• anthropic:claude-opus-4@20250514
• google:gemini-2.0-flash-001
• google:gemini-2.5-pro-preview-03-25
• google:gemini-2.5-flash-preview-04-17
• bedrock:anthropic.claude-3-5-sonnet-20241022-v2:0
• bedrock:anthropic.claude-3-7-sonnet-20250219-v1:0
• bedrock:anthropic.claude-sonnet-4-20250514-v1:0
• bedrock:anthropic.claude-opus-4-20250514-v1:0

架构 | Architecture

应用程序由以下几个模块组成 | The application consists of several modules:

• main.go - 应用程序入口点和服务器设置 | Application entry point and server setup
• config.go - 配置和常量 | Configuration and constants
• models.go - OpenAI和Atlassian API的数据结构 | Data structures for OpenAI and Atlassian APIs
• auth.go - 认证头生成和密码管理 | Authentication header generation and password management
• client.go - 带有重试逻辑和流式支持的HTTP客户端 | HTTP client with retry logic and streaming support
• handlers.go - HTTP请求处理程序 | HTTP request handlers
• transform.go - OpenAI和Atlassian格式之间的数据转换 | Data transformation between OpenAI and Atlassian formats
• db/db.go - 数据库操作 | Database operations
• auth/auth.go - 认证和密码哈希 | Authentication and password hashing
• embed.go - 嵌入静态文件和模板 | Embedded static files and templates

开发 | Development

要启用调试模式以获取详细日志，请在config.go中设置DebugMode = true。

To enable debug mode for verbose logging, set DebugMode = true in config.go.

数据存储 | Data Storage

应用程序使用SQLite数据库（.credentials.db）存储凭证、API令牌和管理员密码。

The application uses an SQLite database (.credentials.db) to store credentials, API tokens, and admin passwords.

许可证 | License

本项目按原样提供，仅供教育和开发目的使用。

This project is provided as-is for educational and development purposes.