qwen-code
by
QwenLM

Description: An open-source AI coding agent that lives in your terminal.

View on GitHub ↗

Summary Information

Updated 2 hours ago
Added to GitGenius on February 18th, 2026
Created on June 26th, 2025
Open Issues & Pull Requests: 617 (+0)
Number of forks: 2,623
Total Stargazers: 25,866 (+0)
Total Subscribers: 137 (+0)

Issue Activity (beta)

Open issues: 554
New in 7 days: 89
Closed in 7 days: 162
Avg open age: 54 days
Stale 30+ days: 409
Stale 90+ days: 215

Recent activity

Opened in 7 days: 68
Closed in 7 days: 66
Comments in 7 days: 194
Events in 7 days: 867

Top labels

  • type/bug (1,506)
  • status/needs-triage (1,097)
  • type/feature-request (652)
  • priority/P2 (387)
  • status/need-information (276)
  • category/core (235)
  • welcome-pr (215)
  • priority/P1 (207)

Repository Insights (GitGenius)

Median issue/PR response: 0.0 hours
Mean response time: 2.5 days
90th percentile: 47.6 hours
Tracked items: 2,705

Most active contributors

Detailed Description

Qwen Code is an open-source AI coding agent designed to operate within terminal environments, built primarily in TypeScript. The project provides developers with an agentic framework that requires minimal setup, featuring capabilities like Auto-Memory, Auto-Skills, SubAgents, Agent Teams, and Model Context Protocol (MCP) support. The framework itself and the underlying Qwen models are both open-source, eliminating vendor lock-in concerns.

The repository supports multiple API protocols including OpenAI, Anthropic, Gemini, and Qwen APIs, with flexibility to integrate third-party providers or local models through Ollama and vLLM. This multi-protocol approach allows users to switch between different language models at runtime. Beyond terminal usage, Qwen Code extends across multiple interfaces including IDE plugins for VS Code, Zed, and JetBrains, a dedicated desktop application for macOS, Windows, and Linux, daemon mode for shared agent sessions over HTTP and Server-Sent Events, TypeScript, Python, and Java SDKs, and instant messaging bot integrations with Telegram, DingTalk, WeChat, and Feishu.

The project operates in several distinct modes to accommodate different workflows. Interactive mode provides a terminal UI with rich rendering and slash commands, headless mode enables scripting and CI/CD integration without graphical output, and daemon mode allows multiple clients to connect to a single shared agent session. The framework includes built-in skills such as code review, batch processing, looping, and bug fixing capabilities.

According to GitGenius activity tracking, the repository demonstrates substantial engagement with a median issue and pull request response latency of zero hours and a mean latency of 58.2 hours across 2674 tracked items. The most prevalent issue categories are bug reports with 1411 tracked instances, items requiring triage with 1015 instances, and feature requests with 597 instances. The core maintenance team includes tanzhenxin with 1952 recorded events, Mingholy with 1692 events, and pomelo-nwu with 1081 events as the most active contributors and triagers.

The repository maintains connections with other significant projects through overlapping contributors, including github/gh-aw, solo-io/gloo, and microsoft/vscode. A notable aspect of the project is that Qwen Code actively uses its own agent and models for internal development, filing issues, submitting pull requests, reviewing code, and running tests. The project documentation is available in multiple languages including Chinese, German, French, Japanese, Russian, and Brazilian Portuguese, reflecting its international scope.

The framework positions itself as feature-complete with Claude Code while offering additional capabilities exclusive to Qwen Code, including open-source models and framework, multi-protocol support, Agent Arena for multi-model comparison, daemon mode functionality, and instant messaging channel integrations. The project originated from Google Gemini CLI v0.8.2 but has evolved into an independent multi-protocol, multi-platform agent framework with specialized integrations for Qwen models and broader ecosystem support.

qwen-code
by
QwenLMQwenLM/qwen-code

Repository Details

Fetching additional details & charts...