Nano Agent
A MCP Server for a small scale engineering agents with multi-provider LLM support.
What is Nano Agent?
Nano Agent is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to mcp server for a small scale engineering agents with multi-provider llm support.
A MCP Server for a small scale engineering agents with multi-provider LLM support.
This server falls under the Coding Agents category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- A MCP Server for a small scale engineering agents with multi
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx nano-agentConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Nano Agent
Nano Agent is a lightweight MCP server that exposes a small set of file-system tools — read, write, edit, list, and inspect — to AI coding agents backed by any major LLM provider including OpenAI, Anthropic, and local Ollama models. It is designed as a minimal, understandable building block for agentic coding workflows, making it easy to swap models, benchmark providers, or compose nano-agents as sub-agents inside larger Claude Code sessions.
Prerequisites
- Python 3.12 or higher
- uv package manager (https://docs.astral.sh/uv/)
- At least one provider API key: OPENAI_API_KEY, ANTHROPIC_API_KEY, or a running Ollama instance
- An MCP-compatible client such as Claude Desktop or Claude Code CLI
- Git to clone the repository
Clone the repository
Clone the nano-agent monorepo from GitHub. The MCP server lives in the apps/nano_agent_mcp_server subdirectory.
git clone https://github.com/disler/nano-agent.git
cd nano-agentSet up environment variables
Copy the sample env files and fill in your LLM provider API keys. You only need keys for the providers you plan to use.
cp ./.env.sample ./.env
cp ./apps/nano_agent_mcp_server/.env.sample ./apps/nano_agent_mcp_server/.env
# Edit .env and add OPENAI_API_KEY and/or ANTHROPIC_API_KEYInstall the MCP server globally
Run the install script to make the 'nano-agent' command available system-wide via uv tool.
cd apps/nano_agent_mcp_server
./scripts/install.sh
# Or manually:
uv tool install -e .Test the file system tools
Verify the five core tools (read_file, write_file, edit_file, list_directory, get_file_info) work correctly using the built-in test command.
cd apps/nano_agent_mcp_server
uv run nano-cli test-toolsAdd to your MCP client
Add the nano-agent server to your .mcp.json or Claude Desktop config. The global install makes 'nano-agent' available as a direct command.
Run a quick agent task from the CLI
Use nano-cli to verify the full agent loop works with your chosen model before connecting through MCP.
cd apps/nano_agent_mcp_server
uv run nano-cli run "List all Python files in the current directory" --model gpt-5-miniNano Agent Examples
Client configuration
MCP client configuration for nano-agent using the globally installed command.
{
"mcpServers": {
"nano-agent": {
"command": "nano-agent",
"args": [],
"env": {
"OPENAI_API_KEY": "your_openai_api_key",
"ANTHROPIC_API_KEY": "your_anthropic_api_key"
}
}
}
}Prompts to try
Example prompts for delegating file-system tasks through nano-agent.
- "Use nano-agent to read the contents of src/main.py and summarize what it does"
- "Ask nano-agent to create a new file called utils.py with a function that parses JSON"
- "Have nano-agent list the directory structure of the project and identify any test files"
- "Use nano-agent with the claude-3-haiku model to edit config.json and update the version field to 2.0.0"
- "Delegate to nano-agent: find all .env files and report their sizes without reading their contents"Troubleshooting Nano Agent
nano-agent command not found after running install.sh
Run 'uv tool list' to confirm the tool was installed. If missing, run 'uv tool install -e .' from inside apps/nano_agent_mcp_server. Ensure uv's tool bin directory is in your PATH (run 'uv tool dir' to find it).
API key errors for OpenAI or Anthropic
Make sure the .env file inside apps/nano_agent_mcp_server/ contains the correct keys, not just the root .env. The server reads from its own directory's .env file at startup.
Ollama model requests fail with connection refused
Start the Ollama service locally before using Ollama models: run 'ollama serve' in a separate terminal. Then specify '--provider ollama' when calling nano-cli to route requests correctly.
Frequently Asked Questions about Nano Agent
What is Nano Agent?
Nano Agent is a Model Context Protocol (MCP) server that mcp server for a small scale engineering agents with multi-provider llm support. It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Nano Agent?
Follow the installation instructions on the Nano Agent GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with Nano Agent?
Nano Agent works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Nano Agent free to use?
Yes, Nano Agent is open source and available under the MIT license. You can use it freely in both personal and commercial projects.
Nano Agent Alternatives — Similar Coding Agents Servers
Looking for alternatives to Nano Agent? Here are other popular coding agents servers you can use with Claude, Cursor, and VS Code.
Dify
★ 142.2kProduction-ready platform for agentic workflow development.
Ruflo
★ 54.0k🌊 The leading agent orchestration platform for Claude. Deploy intelligent multi-agent swarms, coordinate autonomous workflows, and build conversational AI systems. Features enterprise-grade architecture, self-learning swarm intelligence, RAG integrat
Goose
★ 45.7kan open source, extensible AI agent that goes beyond code suggestions - install, execute, edit, and test with any LLM
Antigravity Awesome Skills
★ 38.3kInstallable GitHub library of 1,400+ agentic skills for Claude Code, Cursor, Codex CLI, Gemini CLI, Antigravity, and more. Includes installer CLI, bundles, workflows, and official/community skill collections.
AgentScope
★ 25.5kBuild and run agents you can see, understand and trust.
Serena
★ 24.5kA coding agent toolkit that provides IDE-like semantic code retrieval and editing tools, enabling LLMs to efficiently navigate and modify codebases using symbol-level operations instead of basic file reading and string replacements.
Browse More Coding Agents MCP Servers
Explore all coding agents servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up Nano Agent in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use Nano Agent?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.