Kimi CLI
Kimi CLI is Moonshot AI's terminal coding agent that runs Kimi models against your local repo with tool use, long context, and sandboxed execution.
Kimi CLI is Moonshot AI's command-line coding agent, powered by the Kimi family of models (including Kimi K2 and successors). It runs locally, edits your repo, executes shell commands, and iterates toward a goal — the same agentic coding pattern as Claude Code and Codex CLI. Kimi models are known for very long context windows, which helps on large codebases.
Why it matters
Long context matters for coding agents because every file read, every command output, and every diff consumed ends up in the agent's context. Kimi's long-context advantage means fewer mid-session compactions and fewer "I forgot what I was doing" recoveries when working across many files.
Kimi CLI is one of SpaceSpider's first-class CLIs. Pick it in the wizard and it spawns in its own pane; run it alongside Claude Code, Codex CLI, and Qwen Code in the same grid layout for side-by-side comparison. See cli-kimi.
How it works
Kimi CLI exposes the usual agentic toolset: file read/write, shell execution, search, and optionally MCP servers for external integrations. It authenticates against Moonshot's API with a key and streams responses as the model emits tool use calls. Output is fed back into the context window as observations; the loop continues until the task is done or the user interrupts.
Configuration typically lives under ~/.kimi/ and supports model selection, default approval mode, and custom system prompts.
How it's used
Developers use Kimi CLI for the same workloads as other agentic CLIs:
- Whole-repo refactors that benefit from long context
- Bug fixes across several files
- Test-driven development loops with automatic iteration
- Exploratory investigation via plan mode-style prompts
Running it in SpaceSpider gives you parallel panes so you can compare Kimi's output to Claude or Codex on the same task.
Related terms
- Claude Code — Anthropic's counterpart
- Codex CLI — OpenAI's counterpart
- Qwen Code — the open-weights sibling
- Context window — where Kimi's long window matters
- Agentic coding — the broader category
FAQ
Why pick Kimi over Claude or GPT?
Price-per-token and context length are the main reasons. Kimi models are typically cheaper than frontier Anthropic/OpenAI options and support very long contexts, which suits whole-repo tasks.
Does Kimi CLI support MCP?
Yes, modern versions support the Model Context Protocol, so the same MCP servers you use with Claude Code can be reused with Kimi CLI.
Related terms
- Agentic codingAgentic coding is software development where an LLM-powered agent plans, edits, runs, and verifies code on its own using tools, not just autocomplete.
- AI pair programmingAI pair programming is a collaboration style where an LLM assistant sits alongside you, suggesting code and reviewing changes in real time as you work.
- ANSI escape codesANSI escape codes are control sequences that terminals interpret for colors, cursor movement, and screen clearing — the language of every modern CLI UI.
- Autonomous agentAn autonomous agent is an AI program that perceives, decides, and acts on its own toward a goal — the architecture behind modern coding CLIs.
- CheckpointA checkpoint is a saved snapshot of file state that lets you roll back an AI coding agent's changes to a known-good point.
- Claude CodeClaude Code is Anthropic's official command-line agent that plans, edits, runs, and verifies code across your repo using Claude models and tool use.