home / mcp / float mcp server

Float MCP Server

A community MCP server for float.com.

Installation

Add the following to your MCP client configuration file.

Configuration

{
  "mcpServers": {
    "asachs01-float-mcp": {
      "command": "docker",
      "args": [
        "run",
        "-d",
        "--name",
        "float-mcp",
        "-e",
        "FLOAT_API_KEY=YOUR_API_KEY",
        "-e",
        "LOG_LEVEL=info",
        "-p",
        "3000:3000",
        "float-mcp"
      ],
      "env": {
        "LOG_LEVEL": "info",
        "FLOAT_API_KEY": "YOUR_API_KEY",
        "FLOAT_API_BASE_URL": "https://api.float.com/v3"
      }
    }
  }
}

You can run this MCP server to expose Float’s API as MCP tools, enabling AI assistants to manage Float projects, time tracking, resources, and more through a unified protocol. This server consolidates Float’s capabilities into a single, TypeScript-supported MCP interface with built-in logging, rate limiting, and Docker containerization for easy deployment.

How to use

Leverage the MCP server by connecting your MCP client to the available connection method. The server supports a local runtime (stdio) via Docker and exposes the Float API through the MCP interface. Use the optimized tools to perform common workflows like entity management, project workflows, time tracking, and reporting, while preserving full compatibility with existing client integrations.

How to install

Prerequisites include a modern container runtime and a Float account with API access. The recommended route is to run the server using Docker in a production-like setup.

# Build and run the MCP server using Docker
# 1) Build and run the container
# Replace YOUR_API_KEY with your actual Float API key

# Start the container with environment variables and port exposure

docker run -d \
  --name float-mcp \
  -e FLOAT_API_KEY=YOUR_API_KEY \
  -e LOG_LEVEL=info \
  -p 3000:3000 \
  float-mcp

Configuration and security

Configure your API access via environment variables. The following variables are used by the server: FLOAT_API_KEY for authenticating with Float and LOG_LEVEL for runtime logging. You can adapt these values to your deployment environment.

FLOAT_API_KEY=your_float_api_key_here
LOG_LEVEL=info

Usage patterns and examples

Connect your MCP client to the local runtime at the standard MCP endpoint. Use the prebuilt tools to perform common operations such as creating people, assigning tasks, scheduling time off, generating reports, and more. The four optimized tools cover core functionality while maintaining backward compatibility with legacy tooling.

Troubleshooting and notes

If you encounter authentication or rate-limiting issues, verify that your FLOAT_API_KEY is valid and that the server is reachable on the configured port. The server includes exponential backoff for retry attempts and provides detailed logging when LOG_LEVEL is set to debug.

LOG_LEVEL=debug

Security considerations

Keep your Float API key secure and rotate it periodically. Run the MCP server in a restricted network environment and implement access controls around who can connect MCP clients to the local runtime.

Available tools

manage-entity

Core Entity Management consolidating CRUD operations for people, projects, tasks, departments, roles, accounts, statuses, and more through a unified decision-tree routing.

manage-project-workflow

Handles all project-specific workflow operations such as phases, milestones, project tasks, allocations, dependencies, and bulk operations.

manage-time-tracking

Manages time-related operations including logged time, time off, holidays, approvals, and timesheets with reporting.

generate-report

Analytics and reporting engine with time, project, people utilization, capacity, and budget reports, with advanced filtering and grouping.