M
MCP Directory CURATED DIRECTORY
911 MCP servers indexed
中文
← Back to directory
M

MiniClaude

Community
Not a standard server
Lightweight local AI coding assistant, streamlined from Claude Code

This is not a typical connect-and-use MCP server; read the source repository before using it.

How to install ↓ Compare GitHub source repository ↗
Category
Dev Tools #391 of 438
Stars
★ 171 Popular
Transport
stdio (local process)
Runtime
Bun 1.3.11+
Credentials
API key / credential required
License
Other / unspecified
Last commit
35FMRS · D

Created by community developer txl16095, based on Claude Code, with code size about 20% of the original.

Strongest · Maintenance 10/20 Weakest · Security and permissions 4/20

Reliability
5/20
Security and permissions
4/20
Maintenance
10/20
Documentation
8/20
Setup experience
8/20
Why each score
Reliability 5/20
Evidence shows a complete client application based on Claude Code, but the server manifest is not cached, and the README does not provide a clear MCP server entry point or tool list. The project claims to support MCP protocol, but no implementation details are given. No tests or CI configurations are found. From static review, it cannot be verified that it starts cleanly and completes the MCP handshake. Thus low reliability score.
Security and permissions 4/20
Red line: license is NOASSERTION, and although README claims telemetry removed, the install example contains a placeholder API key (sk-your-api-key-here), which is risky. No obvious malware, but lack of security audit. No security disclosure mechanism. License unclear, so security score limited.
Maintenance 10/20
Repository has 174 stars, 0 open issues, not archived. README suggests active development, but no releases. License NOASSERTION is a legal risk. Moderate maintenance score.
Documentation 8/20
README provides detailed installation and configuration guides, including settings.json and .env examples. Links to official website and docs. But lacks MCP-server-specific documentation, no explanation of how to use as an MCP server. Documentation is good but incomplete.
Setup experience 8/20
README provides clear installation steps using Bun and Git. For MCP server, needs settings.json or .env configuration. No main client config examples provided. Setup is relatively simple but lacks client integration examples.

Static review · not runListed 2026-08-07

Read the FMRS scoring method →

Fit and risk

What it can accessReads local filesWrites / deletes local filesRuns commands or codeUses the networkControls a browser

Best for

  • Developers concerned about privacy and local development
  • Individuals or teams needing a lightweight alternative to Claude Code
  • Developers familiar with Bun and TypeScript

Not for

  • Users needing team collaboration and cloud sync features
  • Enterprises requiring official support and maintenance
  • Windows users not comfortable with command line

Required permissions

  • Read and write project files
  • Execute shell commands (Bash/PowerShell)
  • Access Git repositories
  • Use network services (calling AI APIs)

Risks and side effects

  • When using third-party APIs, be aware of data privacy; code may be sent to external model services
  • Project is not official, there may be security vulnerabilities or incomplete features
  • Depends on Bun runtime, ensure environment stability

Setup

Before you start

Runtime：Bun 1.3.11+

ANTHROPIC_AUTH_TOKEN requiredsecret API key from your chosen model provider (e.g. from the DeepSeek console), set in settings.; treat as a secret.
Other optional settings (2)
ANTHROPIC_BASE_URL optional Anthropic-compatible API endpoint for the model provider, e.g. https://api.deepseek.com/anthropic, set in the env block of ~/.claude/settings..
ANTHROPIC_MODEL optional Model name to use, e.g. deepseek-v4-pro[1m]; can also be set in the top-level model field.

Prerequisites: Bun >= 1.3.11 and Git. Installation and build:

git clone https://github.com/txl16095/MiniClaude.git && cd MiniClaude
bun install
bun run build

Then copy the configuration file:

cp settings.example.json ~/.claude/settings.json

And edit it as needed, for example configuring DeepSeek:

{
  "env": {
    "ANTHROPIC_BASE_URL": "https://api.deepseek.com/anthropic",
    "ANTHROPIC_AUTH_TOKEN": "sk-your-api-key-here",
    "ANTHROPIC_MODEL": "deepseek-v4-pro[1m]"
  },
  "model": "deepseek-v4-pro[1m]"
}

Check that it works

After building, run ./cli to start MiniClaude; if the chat session starts and the /provider command lists available providers, the install and configuration are working.

Troubleshooting

  1. If build fails, ensure Bun >= 1.3.11 and re-run bun install
  2. If AI features do not work, check API key and base URL in settings.json
  3. If environment issues occur, try copying .env.example and using bun run dev for development mode

Things to try

Once connected, you can ask your AI assistant things like:

  • Read this project's code and summarize its structure
  • Refactor the function in this file and write the changes back
  • Run the tests and fix the failing cases
  • Commit the current Git changes

Use cases

Local AI coding assistance without cloud services
Use third-party models (e.g., DeepSeek) for code generation and project understanding
Extend functionality via MCP protocol

Supported clients

Claude Desktop

Listed from the project's documentation, not tested by this site.

Overview

MiniClaude is a lightweight local AI coding assistant derived from Claude Code, which removes cloud services, telemetry, and collaboration code (about 92,000 lines), while retaining core development features such as AI conversation, code generation, file operations, Git, MCP, plugin system, and skills. It supports using various AI models (e.g., DeepSeek) and is configured via settings.json.

Similar servers

Context7 80 · B

Upstash's official server providing up-to-date third-party library docs for AI coding assistants

★ 63.3k · Tools 2 Compare with this →
Compare with other MCP servers

Source revision 561531ea4567 Data synced 2026-10-19 Read the FMRS scoring method