Kimi Code CLI: Moonshot AI's Open-Source Answer to Claude Code

🚨 Moonshot AI just open-sourced a serious Claude Code competitor. It’s called Kimi Code CLI — an AI coding agent that runs directly inside your terminal. And this isn’t just another Claude Code wrapper. Kimi Code CLI can: 🎥 Accept screen recordings and demo videos as input 🧠 Run separate coder, planning and exploration agents in isolated contexts 📝 Create a plan before making changes to your files 🔌 Configure MCP servers conversationally with "/mcp-config" ⚡ Start in milliseconds with no Node.js setup 💻 Work with VS Code, JetBrains, Zed and other ACP-compatible editors 🌐 Search documentation and browse the web 🛠️ Read files, edit code and execute terminal commands The CLI is open source and free to install. You can connect it to Moonshot’s Kimi models or configure another compatible model provider. Kimi K3, Moonshot’s latest flagship model, also supports: → A context window of up to 1 million tokens → Native visual understanding → Long-horizon coding and agentic workflows → Large codebase analysis Claude Code, Codex, Cursor and OpenCode now have another serious competitor. Would you trust an open-source coding agent with your production codebase?

This is a real, recent, and largely accurate release announcement — both the CLI tool and the underlying model check out well against multiple independent sources.

Confirmed accurate:

• Kimi Code CLI is real — an open-source terminal coding agent from Moonshot AI (successor to the earlier kimi-cli), confirmed via Moonshot's own GitHub repo (MoonshotAI/kimi-code), MarkTechPost, and several independent reviews. License sources vary slightly between MIT and Apache-2.0 across the tool's version history — this reflects the project evolving (older kimi-cli was Apache-2.0; the newer TypeScript-based Kimi Code was reported as MIT), not necessarily a contradiction.
• Screen recordings/video input — confirmed as a distinctive, genuinely unusual feature; multiple sources specifically flag video input as one of the two standout features that differentiate it from similar tools.
• Separate coder, planning, and exploration subagents in isolated contexts — confirmed directly from Moonshot's own GitHub README: "Subagents for focused, parallel work. Dispatch built-in coder, explore, and plan subagents in isolated contexts."
• Creates a plan before making changes — consistent with its described "autonomous planning" and step-by-step feedback-driven execution model.
• Conversational MCP configuration via /mcp-config — confirmed verbatim from Moonshot's own repo description.
• Fast startup, no Node.js setup — confirmed for the current version; later coverage describes it as "a single binary with no Node.js dependency" and a "Rust-backed TUI that launches in milliseconds" (an earlier TypeScript/npm-based version did exist, but the current shipped version matches the post's claim).
• ACP compatibility with Zed, JetBrains, and other editors — confirmed directly.
• Web search/browsing, file read/edit, terminal command execution — all confirmed core capabilities.
• Free and open source, connects to Kimi models or other compatible providers — confirmed.
• Kimi K3's specs: 1M-token context window, native visual understanding (text/image/video), and positioning for long-horizon coding/agentic workflows and large codebase analysis — all directly confirmed across numerous independent sources (VentureBeat, DataCamp, Tom's Hardware, Moonshot's own platform docs).

Worth noting as a nuance:

• Kimi K3 itself is described by outlets as "open-weight" under a bespoke Kimi K3 License, not OSI-approved open source — the post doesn't explicitly call K3 "open source" (it calls the CLI open source, correctly), so this isn't a factual error, but it's worth knowing the model's licensing is more restrictive than a standard open-source license.
• K3 reportedly carries a notably high hallucination rate (one source cites 51%) and significant hardware/serving costs — context the post doesn't mention, though it wasn't claiming otherwise.

Verdict: This is an accurate, well-corroborated announcement. Every specific technical claim about the CLI and the model checks out against Moonshot's own materials and independent tech coverage.