在当今社会，代码协作变得越来越重要，特别是在分布式团队和多任务场景下。然而，现有的协作工具大多集中在代码管理和文档协作，缺乏强大的多智能体人工智能（AI）辅助功能。xintaofei/codeg 的出现正是为了填补这一空白，通过聚合来自 Claude Code, Codex, OpenCode, Pi, Grok Build 等多个 AI 编码助手的会话，以实现更高效和智能的多代理协作开发环境。
核心技术揭秘
技术架构分析
xintaofei/codeg 搭建了一个覆盖多平台的协作环境，支持桌面应用程序、自托管服务器，以及 Docker 容器化部署。其核心在于利用先进的 AI 工具集成，提供了强大的协作平台。
graph TD
A[用户请求] --> B{"参数校验"}
B -->|"通过"| C[执行业务]
C --> D[数据处理]
D --> E{"环境检测"}
E -->|"桌面客户端"| F[桌面应用程序]
E -->|"服务器托管"| G[自托管服务器]
E -->|"容器化部署"| H[Docker 容器]
Collaborative multi-agent AI coding workspace: aggregate sessions from Claude Code, Codex, OpenCode, Pi, Grok Build, etc. Desktop app, self-hosted server, or Docker.
Thanks to Compshare for sponsoring this project! Compshare is UCloud's AI cloud platform, offering cost-effective monthly and pay-as-you-go agent Plan subscriptions for Chinese models, starting at just ¥49/month. It also provides stable officially-proxied access to overseas models. Supports Claude Code, Codex, and API integrations. Enterprise-ready with high concurrency, 24/7 technical support, and self-service invoicing. Users who sign up via this link receive ¥5 in free platform credits!
Thanks to 合租巴士 for sponsoring this project! 合租巴士 is a reliable and efficient AI relay platform, offering highly stable relay for mainstream models such as Codex and Claude Code. Top-ups are credited at a transparent 1:1 ratio, with Codex rate subsidies as low as 0.08. Join the group via the official website to get $5 in trial credit.
Thanks to APIMart for sponsoring this project! APIMart is a low-cost API platform for AI image & video generation — GPT-Image-2 from $0.006/image, 160+ images per dollar. One async API covers both image and video: submit a task, get an ID, fetch results via polling or callback. Batch tens of thousands of images without timeouts, switch models without changing code. Pay-as-you-go with no monthly fee — sign up here to get started.
UCloud ·星图AstraFlow
UCloud's AstraFlow large-model platform gives you one-click access to 200+ models: leading open-source models such as Kimi K3, DeepSeek V4/V3, Qwen 3, GLM5.2, and happyhorse are built in — no training of your own required, ready to use out of the box.
Sign up with your email through the link above, complete real-name verification, and claim ¥50 in compute credits.
Thanks to Sidrune AI for sponsoring this project! Sidrune AI gives you access to GPT, Claude, and other major AI models through a single API. New users who sign up through our dedicated link with promo code SDRCODEG get $3 in API credits.
Thanks to BeeAPI for sponsoring this project! BeeAPI is a professional multi-model AI API relay and aggregation platform that brings together many providers and a wide range of mainstream AI models. It supports cross-vendor price comparison, smart routing across groups, and unified API access and billing, helping you call AI services more flexibly, reliably, and at a lower cost.
Repository Panel — issues, pull requests, checks, and merges for GitHub, GitLab, Gitea, and Forgejo; any issue or PR can become a to-do
Automations — save a configured composer and run it headlessly, on a cron schedule or on demand
Token Usage — trends, cache hit rate, an activity heatmap, usage by folder, agent, model, and session, and what it would cost at API list prices
Fork & Steer — fork a conversation from a finished reply, or send a message into a turn that is still running, on agents that support it
Skills & MCP — skill packs per agent, a local MCP scan, and registry search and install
Project Boot — scaffold a new project visually, with a live preview
Notifications — desktop notifications, sounds, and a flashing taskbar button or bouncing Dock icon when a turn ends or an agent needs you
Make It Yours — twelve themes you can recolor token by token, wallpapers, corner radius, chat width, custom CSS, and a desktop pet — in any of ten interface languages
Backup & Sync — encrypted backups, and configuration sync between machines through a file or your own WebDAV server
URL Scheme — codeg://session/<id> opens a conversation from another app (desktop)
And more — nearly every release adds something; the release notes are the full list
🤖 Supported Agents
Codeg talks to every agent over the Agent Client Protocol, so each one gets the same structured interface. Twenty come built in, and Codeg installs, pins, and updates most of them for you:
Not on the list? Pick any agent from the public ACP registry or paste its distribution JSON — Codeg installs it, checks that it launches, and treats it like a built-in. Runtime requirements and where each agent keeps its sessions: Supported Agents.
📦 Install
Desktop — installers for macOS, Windows, and Linux are on Releases; setup steps in Installation.
Server — run Codeg headless and reach it from any browser. On Linux or macOS:
docker run -d -p 3080:3080 -v codeg-data:/data ghcr.io/spacering-net/codeg:latest
Either way, open http://<host>:3080 in a browser and enter the access token the server prints when it starts (docker logs shows it for the container), or set CODEG_TOKEN to choose your own.
Mobile — install the iOS app or the Android APK, then point it at the Web Service of your desktop app or at your own codeg-server: URL, token, done. Pairing steps in Mobile apps.
Compose, prebuilt binaries, source builds, and in-place updates are covered in Deployment; environment variables in Configuration.
👥 Community & Support
WeChat — scan the QR code to join our group for discussions, feedback, and updates:
Issues — found a bug or missing a feature? Open an issue.
Privacy — local-first: parsing, storage, and project operations stay on your machine, and web and server modes are guarded by token authentication. Details in Privacy & Security.
LinuxDO — thanks to the LinuxDO community for their support.
Claude Code
Codex
Gemini CLI
OpenClaw
Kimi Code
MiniMax Code
Pi
DeepSeek Harness
Kiro CLI
ZCode
TRAE CLI
Command Code
评论