MoonshotAI/kimi-code

Kimi Code CLI — The Starting Point for Next-Gen Agents

View on GitHub ↗Jump to charts ↓Open shareable report

Summary Information

Updated 55 minutes ago
Added to GitGenius on July 22nd, 2026
Created on May 22nd, 2026
Open Issues & Pull Requests: 1,342 (-1)
GitHub issues: Enabled
Number of forks: 1,184
Total Stargazers: 7,337 (+0)
Total Subscribers: 33 (+0)

Repository Insights (GitGenius)

Median issue/PR response: 17.1 hours
Mean response time: 5.6 days
90th percentile: 15.7 days
Tracked items: 680

How this project is maintained

Around half of the issues opened in the past year never receive a reply. 98% of open issues come from outside the core team, so the backlog reflects real-world use rather than internal planning. Almost all tracked open issues have seen activity in the last three months. Only 9% of issues opened in the past year have been closed. Three people close 59% of everything that gets resolved.

Charts & Analytics

Fetching additional details & charts...

Issue Activity (beta)

Open issues: 916
New in 7 days: 71
Closed in 7 days: 14
Avg open age: 12 days
Stale 30+ days: 575
Stale 90+ days: 78

Recent activity

Opened in 7 days: 64
Closed in 7 days: 10
Comments in 7 days: 54
Events in 7 days: 95

Top labels

  • bug (347)
  • enhancement (211)

Detailed Description

Kimi Code CLI is an AI coding agent designed to run directly in the terminal, built by Moonshot AI and written in TypeScript. The tool functions as an autonomous agent capable of reading and editing code, executing shell commands, searching files, fetching web pages, and making decisions based on feedback received during execution. It integrates natively with Moonshot AI's Kimi models while remaining configurable to work with other compatible AI providers.

The installation process is streamlined for accessibility, requiring only a single command with no Node.js dependency for end users, though development requires Node.js 24.15.0 or later and pnpm 10.33.0. The tool supports macOS, Linux, and Windows environments, with Windows users needing Git for Windows installed beforehand.

Key architectural features include single-binary distribution that eliminates setup friction, millisecond-level startup times for the terminal UI, and a purpose-built interface optimized for extended agent sessions. The tool supports video input capabilities, allowing users to provide screen recordings or demo clips that the agent can analyze and convert into actionable outputs like code or visual transformations. Configuration of Model Context Protocol servers can be handled conversationally through the /mcp-config command, avoiding manual JSON editing. The plugin ecosystem enables installation of skills, MCP servers, and data sources from the marketplace or GitHub repositories, with trust levels surfaced during installation.

The platform includes subagent functionality with built-in coder, explore, and plan agents that can work in parallel within isolated contexts while maintaining a clean main conversation thread. Lifecycle hooks allow users to run local commands at critical points for gating risky operations, auditing decisions, triggering notifications, or connecting to external automation systems. Integration with editors and IDEs is supported through the Agent Client Protocol, enabling Zed, JetBrains, and other compatible clients to drive Kimi Code CLI sessions directly through the kimi acp subcommand without requiring separate authentication.

The repository maintains connections with related projects including hmbown/codewhale, hmbown/deepseek-tui, and moonshotai/kimi-cli through overlapping contributor networks. The project is released under the MIT License and acknowledges the pi-tui library as a foundation for its terminal user interface implementation.