kimi-code-mcp
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@kimi-code-mcp search the web for MCP server tutorials"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
kimi-code-mcp
MCP server wrapping Kimi Code CLI (kimi-k2.5) — 14 tools for filesystem, shell, web, and agent operations.
Prerequisites
Node.js >= 18
Kimi CLI installed and authenticated
Related MCP server: emcp
Usage with Claude Desktop
Add to your Claude Desktop configuration:
{
"mcpServers": {
"kimi": {
"command": "npx",
"args": ["-y", "kimi-code-mcp@latest"]
}
}
}Usage with VS Code
Install the Kimi Code MCP server in VS Code:
Or add to .vscode/mcp.json:
{
"servers": {
"kimi": {
"command": "npx",
"args": ["-y", "kimi-code-mcp@latest"]
}
}
}Tools
Tool | Description |
| Read a text file (up to 1000 lines) |
| Analyze images and videos |
| Create or overwrite a file |
| Find-and-replace edit in a file |
| Find files matching a glob pattern |
| Search for regex patterns in files |
| Execute shell commands |
| Search the web (up to 20 results) |
| Fetch and extract webpage content |
| Autonomous agent for complex tasks |
| Extended reasoning and analysis |
| Code review (bugs, security, perf, style) |
| Research with 256K context window |
| Generate or fix tests with edge cases |
Environment Variables
Variable | Description | Default |
| Absolute path to kimi binary |
|
| Enable debug logging to stderr | unset |
Development
git clone https://github.com/userfrm/kimi-code-mcp.git
cd kimi-code-mcp
npm install
npm run build
npm testLicense
MIT
Available Tools
14 toolskimi_agentC
Full autonomous agent for complex multi-step tasks.
Inputs: prompt (required), timeout, workFolder.
kimi_edit_fileA
Edit a file by replacing an exact string with new content.
Inputs: path (required), new_string (required), old_string (required), workFolder.
kimi_fetch_urlB
Fetch and extract text content from a webpage.
Inputs: url (required), prompt.
kimi_globA
Find files matching a glob pattern.
Inputs: pattern (required), path, workFolder.
kimi_grepB
Search for regex patterns in files using ripgrep.
Inputs: pattern (required), path, include, workFolder.
kimi_read_fileA
Read a text file. Returns file content up to 1000 lines.
Inputs: path (required), limit, offset, workFolder.
kimi_read_mediaB
Analyze images and videos using multimodal capabilities.
Inputs: path (required), prompt, workFolder.
kimi_researchD
Research and analysis with 256K context window.
Inputs: question (required), context, workFolder.
kimi_reviewB
Code review analyzing bugs, security, performance, and style.
Inputs: code_or_path (required), focus, workFolder.
kimi_shellB
Execute shell commands and return output.
Inputs: command (required), timeout, workFolder.
kimi_testC
Generate or fix comprehensive tests with edge cases.
Inputs: target (required), workFolder, instructions.
kimi_thinkA
Extended reasoning and analysis without taking actions.
Inputs: problem (required), context.
kimi_web_searchB
Search the web and return up to 20 results.
Inputs: query (required), include_content.
kimi_write_fileA
Create or overwrite a file with specified content.
Inputs: path (required), content (required), workFolder.
Tool Schema Changelog
Recent tool additions, removals, and schema changes observed during successful MCP inspections.
Loading tool schema changelog…
TDQS
Scored across 14 tools
Most tools have clearly distinct purposes: file operations (write/read/edit), search (glob/grep), web (search/fetch), and media analysis are unambiguous. Some potential overlap exists between agent, think, review, and research, but their descriptions differentiate them sufficiently for agent selection.
All tools share the consistent 'kimi_' prefix, which provides strong brand consistency. However, the suffix style varies: some are verb_noun (write_file, read_media), while others are single verbs or nouns (glob, shell, think, review). This is a minor deviation from a strict verb_noun pattern but remains predictable.
With 14 tools, the server is well within the ideal 3-15 range. Each tool covers a distinct capability area for a coding assistant, and none feel redundant or superfluous. The count feels appropriately scoped for the server's purpose.
The tool surface covers core coding needs: file manipulation, search, shell execution, web access, and specialized tasks like review, test, and research. Minor gaps exist (e.g., no explicit delete or rename file tool), but these can be handled by the agent or shell tools, so agents can work around them.
Maintenance
Related MCP Connectors
Nifty's MCP server — exposes tasks, projects, messages, and files as tools for AI agents.
MCP server for progressive tool usage at any scale (see https://klavis.ai)
Remote MCP server exposing your Node Command fleet (files, shell, processes, browser) to AI hosts.
Shared memory MCP server for Claude Code, Codex, Gemini CLI and Cursor. Cited, editable.
Related MCP Servers
- AlicenseAqualityCmaintenanceMCP server that turns Kimi K2.6 Turbo into an agentic coding assistant with tools for file operations, shell commands, and code search.429 PyPI2MIT
- FlicenseNot gradedqualityDmaintenanceMCP server providing filesystem operations, shell execution, and web search capabilities.-
- AlicenseNot gradedqualityFmaintenanceMCP server that runs Kimi K2.7-code as an autonomous coding agent inside a bubblewrap-sandboxed git worktree, returning git diffs of changes for review.MIT
- AlicenseNot gradedqualityDmaintenanceMCP server that provides file operations and Moonshot API-powered tools (reasoning, code review, testing, research, web search, agent) for Kimi K2.5, requiring only a Moonshot API key.199 npm1MIT