Codeg

English | 简体中文 | 繁體中文 | 日本語 | 한국어 | Español | Deutsch | Français | Português | العربية

Codeg (Code Generation) is an enterprise-grade multi-agent coding workspace. It unifies local AI coding agents (Claude Code, Codex CLI, OpenCode, Gemini CLI, OpenClaw, Cline, etc.) in a desktop app, standalone server, or Docker container — enabling remote development from any browser — with conversation aggregation, parallel git worktree development, MCP/Skills management, chat channel interactions (Telegram, Lark, iLink, etc.), and integrated Git/file/terminal workflows.

Agents	MCP	Skills	Version Control	Web Service

• Unified multi-agent workspace in the same project
• Local conversation ingestion with structured rendering
• Parallel development with built-in git worktree flows
• Project Boot — visually scaffold new projects with live preview
• Chat Channels — connect Telegram, Lark (Feishu), iLink (Weixin) and more to your coding agents for real-time notifications, full session interaction, and remote task control
• MCP management (local scan + registry search/install)
• Skills management (global and project scope)
• Git remote account management (GitHub and other Git servers)
• Web service mode — access Codeg from any browser for remote work
• Standalone server deployment — run codeg-server on any Linux/macOS server, access via browser
• Docker support — docker compose up or docker run, with custom token, port, and volume mounts for data persistence and project directories
• Integrated engineering loop (file tree, diff, git changes, commit, terminal)

Create new projects visually with a split-pane interface: configure on the left, preview in real time on the right.

• Visual Configuration — pick style, color theme, icon library, font, border radius, and more from dropdowns; the preview iframe updates instantly
• Live Preview — see your chosen look & feel rendered in real time before creating anything
• One-Click Scaffolding — hit "Create Project" and the launcher runs shadcn init with your preset, framework template (Next.js / Vite / React Router / Astro / Laravel), and package manager of choice (pnpm / npm / yarn / bun)
• Package Manager Detection — automatically checks which package managers are installed and shows their versions
• Seamless Integration — the newly created project opens in Codeg's workspace right away

Currently supports shadcn/ui project scaffolding, with a tab-based design ready for more project types in the future.

Connect your favorite messaging apps — Telegram, Lark (Feishu), iLink (Weixin), and more — to your AI coding agents. Create tasks, send follow-up messages, approve permissions, resume sessions, and monitor activity — all from your chat app. Receive real-time agent responses with tool-call details, permission prompts, and completion summaries without ever opening a browser.

More channels (Discord, Slack, DingTalk, etc.) are planned for future releases.

• Session Interaction — run full agent sessions from your chat app: /folder to pick a project, /agent to choose an agent, /task <description> to start a task, then send follow-up messages as plain text. /resume to continue a previous session, /cancel to abort, /sessions to list active sessions
• Permission Control — agents can request tool-execution permissions in-chat; respond with /approve (or /approve always for auto-approve) and /deny without switching context
• Event Notifications — agent turn completions, tool calls, and errors are pushed to all enabled channels in real time with rich formatting
• Query Commands — /search <keyword>, /today, /status, /help for quick lookups; configurable command prefix
• Daily Reports — automated daily summary at a scheduled time, including conversation counts, agent-type breakdown, and project activity
• Multi-Language — message templates in 10 languages (English, Simplified/Traditional Chinese, Japanese, Korean, Spanish, German, French, Portuguese, Arabic)
• Secure Credentials — tokens stored in the OS keyring, never exposed in config files or logs
• Rich Messages — Markdown formatting for Telegram, card-based layout for Lark; plain-text fallback for all platforms

1. Create a channel in Settings → Chat Channels (choose Telegram, Lark, or iLink)
2. Enter your bot token (Telegram), app credentials (Lark), or scan QR code to log in (iLink) — stored securely in the OS keyring
3. Configure event filters and optional daily report schedule
4. Connect — messages start flowing as agents emit events

Note: environment variables take precedence over fallback paths.

• Node.js >=22 (recommended)
• pnpm >=10
• Rust stable (2021 edition)
• Tauri 2 build dependencies (desktop mode only)

Linux (Debian/Ubuntu) example:

sudo apt-get update
sudo apt-get install -y \
  libwebkit2gtk-4.1-dev \
  libayatana-appindicator3-dev \
  librsvg2-dev \
  patchelf

pnpm install

# Frontend static export to out/
pnpm build

# Full desktop app (Tauri + Next.js)
pnpm tauri dev

# Frontend only
pnpm dev

# Desktop build
pnpm tauri build

# Standalone server (no Tauri/GUI required)
pnpm server:dev

# Build server release binary
pnpm server:build

# Lint
pnpm eslint .

# Rust checks (run in src-tauri/)
cargo check
cargo clippy
cargo build

Codeg can run as a standalone web server without a desktop environment.

curl -fsSL https://raw.githubusercontent.com/xintaofei/codeg/main/install.sh | bash

Install a specific version or to a custom directory:

curl -fsSL https://raw.githubusercontent.com/xintaofei/codeg/main/install.sh | bash -s -- --version v0.5.2 --dir ~/.local/bin

Then run:

codeg-server

irm https://raw.githubusercontent.com/xintaofei/codeg/main/install.ps1 | iex

Or install a specific version:

.\install.ps1 -Version v0.5.2

Pre-built binaries (with bundled web assets) are available on the Releases page:

# Example: download, extract, and run
tar xzf codeg-server-linux-x64.tar.gz
cd codeg-server-linux-x64
CODEG_STATIC_DIR=./web ./codeg-server

# Using Docker Compose (recommended)
docker compose up -d

# Or run directly with Docker
docker run -d -p 3080:3080 -v codeg-data:/data ghcr.io/xintaofei/codeg:latest

# With custom token and project directory mounted
docker run -d -p 3080:3080 \
  -v codeg-data:/data \
  -v /path/to/projects:/projects \
  -e CODEG_TOKEN=your-secret-token \
  ghcr.io/xintaofei/codeg:latest

The Docker image uses a multi-stage build (Node.js + Rust → slim Debian runtime) and includes git and ssh for repository operations. Data is persisted in the /data volume. You can optionally mount project directories to access local repos from within the container.

pnpm install && pnpm build          # build frontend
cd src-tauri
cargo build --release --bin codeg-server --no-default-features
CODEG_STATIC_DIR=../out ./target/release/codeg-server

Environment variables:

Next.js 16 (Static Export) + React 19
        |
        | invoke() (desktop) / fetch() + WebSocket (web)
        v
  ┌─────────────────────────┐
  │   Transport Abstraction  │
  │  (Tauri IPC or HTTP/WS) │
  └─────────────────────────┘
        |
        v
┌─── Tauri Desktop ───┐    ┌─── codeg-server ───┐
│  Tauri 2 Commands    │    │  Axum HTTP + WS    │
│  (window management) │    │  (standalone mode)  │
└──────────┬───────────┘    └──────────┬──────────┘
           └──────────┬───────────────┘
                      v
            Shared Rust Core
              |- AppState
              |- ACP Manager
              |- Parsers (conversation ingestion)
              |- Chat Channels
              |- Git / File Tree / Terminal
              |- MCP marketplace + config
              |- SeaORM + SQLite
                      |
              ┌───────┼───────┐
              v       v       v
  Local Filesystem  Git   Chat Channels
    / Git Repos    Repos  (Telegram, Lark, iLink)

• Frontend uses static export (output: "export")
• No Next.js dynamic routes ([param]); use query params instead
• Tauri command params: frontend camelCase, Rust snake_case
• TypeScript strict mode

• Local-first by default for parsing, storage, and project operations
• Network access happens only on user-triggered actions
• System proxy support for enterprise environments
• Web service mode uses token-based authentication

• LinuxDO — the community where it all began
• ACP — the Agent Client Protocol (ACP) is the foundation that enables Codeg to connect with multiple agents

Apache-2.0. See LICENSE.