Reviewer

Reviewer MCP is an MCP server that delivers AI-powered development workflow tools through a configurable, extensible interface. It supports specification generation and review, code review, and project management tasks, making it easier to design, validate, and improve software projects using local or remote AI providers.

How to use

You run the Reviewer MCP server as a local process and connect to it from your MCP client or editor. The server exposes tools for generating and reviewing specifications, reviewing code, and running tests and linters. You can configure multiple providers and point your client at the local stdio server to start a streamlined AI-assisted workflow.

How to install

Prerequisites: you need Node.js and npm installed on your machine.

Install dependencies and build the server:

npm install
npm run build

Configuration and usage notes

Configure how you want Reviewer MCP to run and which AI provider to use by setting environment variables and project-level configuration files.

Environment variables we reference include the AI provider, API keys, and model names. You can structure these in a local .env file and use an example as a guide when configuring your environment.

If you want to run Reviewer MCP with Claude Desktop, connect via a local stdio MCP entry that starts the server process and forwards environment variables like your API key.

If you want to run Reviewer MCP with Ollama (local models), you can configure the system to use a local Ollama instance and, if needed, point your client to a base URL.

Example configurations

{
  "mcpServers": {
    "reviewer": {
      "type": "stdio",
      "name": "reviewer",
      "command": "node",
      "args": ["/path/to/reviewer-mcp/dist/index.js"],
      "env": {
        "OPENAI_API_KEY": "your-api-key-here"
      }
    }
  },
  "envVars": [
    {"name": "AI_PROVIDER", "description": "AI provider to use (openai or ollama)", "required": true, "example": "openai"},
    {"name": "OPENAI_API_KEY", "description": "OpenAI API key for the openai provider", "required": false, "example": "sk-..."},
    {"name": "OPENAI_MODEL", "description": "OpenAI model name to use", "required": false, "example": "gpt-4"},
    {"name": "OLLAMA_BASE_URL", "description": "Base URL for local Ollama instance", "required": false, "example": "http://localhost:11434"},
    {"name": "OLLAMA_MODEL", "description": "Ollama model name to use", "required": false, "example": "llama2"}
  ]
}

Using with Claude Desktop

{
  "mcpServers": {
    "reviewer": {
      "command": "node",
      "args": ["/path/to/reviewer-mcp/dist/index.js"],
      "env": {
        "OPENAI_API_KEY": "your-api-key-here"
      }
    }
  }
}

Using with Ollama

If you want to use local models via Ollama, install Ollama, pull a model, and configure Reviewer MCP to connect to your local Ollama instance. You can use a base URL like http://localhost:11434 and a model such as llama2 or codellama.

# Install Ollama and pull a model
# Visit https://ollama.ai for details

ollama pull llama2

# After pulling the model, configure your environment
# AI_PROVIDER=ollama
# OLLAMA_BASE_URL=http://localhost:11434
# OLLAMA_MODEL=llama2

Available tools and how they work

Reviewer MCP includes a set of standardized tools you can invoke from your MCP client. You can generate specifications from prompts, review existing specifications, review code changes, run tests, and run linters with structured output.

Available tools

generate_spec

Create a technical specification document from a description or context, with optional output format (markdown or structured).

review_spec

Assess a specification document for completeness and quality, providing targeted feedback.

review_code

Analyze code changes (diff) and provide feedback focused on security, performance, style, or logic.

run_tests

Execute project tests with standardized formatting suitable for LLMs, using a configurable test command.

run_linter

Run a linter and produce a structured, machine-readable report.