Harness
MCP server giving AI clients access to the Harness delivery and cloud cost APIs
- Pricing
- Open source
- Platforms
- CLI, API, self-hosted
- Runs
- Locally
- Version
- 0.1.0
- Listing updated
- today
Similar MCP servers
About Harness
Harness is an open-source MCP server for developer tools. It runs locally on your machine and is started by your AI client.
The README describes this as a Model Context Protocol server that connects Harness, a continuous delivery and cloud cost management platform, to GenAI applications. It runs over stdio and is distributed as a Python package installable with pip, with an option to install from a cloned repository.
Access is controlled by a Harness API key supplied through an environment variable, with an optional variable for the API URL. The README mentions multiple authentication methods, rate limiting and retry handling, and async support, but it does not list the individual API endpoints or MCP tools the server exposes.
It suits developers and platform teams who want an AI assistant to query or act on their Harness account through MCP. The project is at an early 0.1.0 release and is published under the MIT license.
Install
Run in a terminal
claude mcp add harness-mcp-server --env HARNESS_API_KEY='<HARNESS_API_KEY>' -- uvx harness-mcp-serverAdd to claude_desktop_config.json
{
"mcpServers": {
"harness-mcp-server": {
"command": "uvx",
"args": [
"harness-mcp-server"
],
"env": {
"HARNESS_API_KEY": "<HARNESS_API_KEY>"
}
}
}
}Add to ~/.cursor/mcp.json
{
"mcpServers": {
"harness-mcp-server": {
"command": "uvx",
"args": [
"harness-mcp-server"
],
"env": {
"HARNESS_API_KEY": "<HARNESS_API_KEY>"
}
}
}
}Add to .vscode/mcp.json
{
"servers": {
"harness-mcp-server": {
"type": "stdio",
"command": "uvx",
"args": [
"harness-mcp-server"
],
"env": {
"HARNESS_API_KEY": "<HARNESS_API_KEY>"
}
}
}
}Add to ~/.codeium/windsurf/mcp_config.json
{
"mcpServers": {
"harness-mcp-server": {
"command": "uvx",
"args": [
"harness-mcp-server"
],
"env": {
"HARNESS_API_KEY": "<HARNESS_API_KEY>"
}
}
}
}Add to your client's MCP settings (most use this shape)
{
"mcpServers": {
"harness-mcp-server": {
"command": "uvx",
"args": [
"harness-mcp-server"
],
"env": {
"HARNESS_API_KEY": "<HARNESS_API_KEY>"
}
}
}
}Using several servers? Build one config file with the MCP config generator.
Configuration
HARNESS_API_KEYAPI Key for HARNESSRequiredHARNESS_API_URLAPI URL for HARNESS
What it does
- Stdio transport
- Runs as a local MCP server over standard input and output for compatible AI clients.
- Harness API access
- Wraps Harness API calls so an AI client can work with a Harness account.
- Configurable authentication
- Reads a required API key and an optional API URL from environment variables.
- Rate limiting and retries
- The README cites built-in rate limiting and retry logic for API calls.
- Async support
- The README says the server supports asynchronous operation.
Good for
- Asking an AI assistant about Harness pipelines and deployments from within a chat or IDE
- Automating routine Harness API queries through an MCP-enabled agent
- Reviewing cloud cost information from Harness via an AI client
Pricing
Open source
Free and open source under the MIT License.
Harness FAQ
How is the server installed?
The README gives pip install harness-mcp-server, or installing from the GitHub source with pip install -e.
What credentials does it need?
The registry entry lists HARNESS_API_KEY as a required secret and HARNESS_API_URL as an optional variable.
What license does it use?
The README states the project is released under the MIT License.
Is Harness free?
Yes. Harness is open source, so you can use and self-host it for free.
How do I add Harness to Claude Code?
Run this command in your terminal: claude mcp add harness-mcp-server --env HARNESS_API_KEY='<HARNESS_API_KEY>' -- uvx harness-mcp-server
What are the best alternatives to Harness?
The closest alternatives to Harness are Uipath, Rundeck, fastapi_mcp, MCP-Bridge and MCP Foundry — Build, Test & Publish MCP Servers.
Where this comes from
- Description: written from the project's README and its MCP Registry entry.
- Install steps: from the official MCP Registry entry.

