ras-commander-mcp
Enables querying HEC-RAS project information, geometry elements, plan results, and compute messages through natural language, allowing interaction with hydraulic modeling projects.
README
HEC-RAS Model Context Protocol (MCP) Server
📖 Documentation: rascommander.info/mcp · full library docs at rascommander.info/ras
<div align="center"> <img src="ras_commander_mcp_logo.svg" alt="RAS Commander MCP Logo" width="70%"> </div>
The RAS Commander MCP (Model Context Protocol) server provides tools for querying HEC-RAS project information using the ras-commander library. This allows Claude Desktop to interact with HEC-RAS hydraulic modeling projects.
RAS Commander MCP is an open-source, LLM-forward H&H automation tool provided under MIT license by CLB Engineering Corporation. This is third-party software and is not made by or endorsed by the U.S. Army Corps of Engineers (USACE) Hydrologic Engineering Center (HEC).
For a demonstration of CLB's H&H automation services, contact us at info@clbengineering.com
When to use the MCP vs. ras-commander directly
| Need | Best fit |
|---|---|
| Richest, most capable workflow | Use a local coding agent (Claude Code or Codex) directly with the ras-commander library and repository. This gives access to the full Python API surface: scripting, notebooks, batch automation, custom analysis, and geometry/results extraction beyond what any fixed tool set exposes. |
| Simple/basic MCP queries | Use this MCP server for a deliberately limited subset of ras-commander: project info, plan results, HDF structure, and geometry element listings. It is especially useful from Claude Desktop and other MCP clients where running a full agent against the repo is not practical. |
Rule of thumb: if your workflow needs custom logic, iteration, writing files, or the broader API, use ras-commander directly with a local agent.
Features
- Query comprehensive HEC-RAS project information (plans, geometries, flows, boundaries)
- List HEC-RAS geometry elements including 1D rivers/reaches, cross sections, 2D reference lines, boundary lines, breaklines, structures, and mesh areas
- Extract detailed plan results including unsteady simulation info and runtime metrics
- Query computed plan results directly, including profile lists, plan summary metrics, 2D mesh cells, and 1D cross sections
- Explore HDF file structures and extract computation messages
- Support for multiple HEC-RAS versions (6.5, 6.6, etc.)
- Formatted text output suitable for LLM interaction
- Error handling with helpful diagnostics
Future Features
-
Summary Results at Boundaries (Max WSE and Max Flow)
-
Also Structures (list them all with max wse and max flow)
-
Detailed XSEC results table (by river/reach) for debugging
-
HEC-RAS Documentation Search Capability (return relevant confluence document links via RAG or deep research)
Prerequisites
- HEC-RAS Installation: HEC-RAS must be installed on your system (default expects version 6.6)
- Python: Python 3.10+
- Claude Desktop: For MCP integration
- uv: Python package manager (recommended)
Installation
This MCP server uses the "uv" library to handle python virtual environments. Once this is installed, the Claude Desktop configuration will handle the package installation and running the MCP server locally whenever Claude Desktop (or Claude Code) are started.
Using uv (Recommended)
- Install uv if you haven't already:
# Windows (PowerShell)
powershell -ExecutionPolicy ByPass -c "irm https://astral.sh/uv/install.ps1 | iex"
# macOS/Linux
curl -LsSf https://astral.sh/uv/install.sh | sh
- Clone this repository:
git clone <repository-url>
cd ras-commander-mcp
- The dependencies will be automatically installed when you run the server with uvx (see Configuration below).
Configuration
Claude Desktop Integration via uvx/pip package (Default, Recommended)
Add the following to your Claude Desktop configuration file (claude_desktop_config.json):
{
"mcpServers": {
"hecras": {
"command": "uvx",
"args": ["ras-commander-mcp"],
"env": {
"HECRAS_VERSION": "6.6"
}
}
}
}
This will install the pip package and set up a virtual environment through uvx to run the MCP server.
Alternate Install 1: From Git Repository (For latest and greatest version)
{
"mcpServers": {
"hecras": {
"command": "uvx",
"args": [
"--from", "ras-commander-mcp@git+https://github.com/gpt-cmdr/ras-commander-mcp.git",
"ras-commander-mcp"
],
"env": {
"HECRAS_VERSION": "6.6"
}
}
}
}
Alternate Install 2: Local Development Installation
If you've cloned the repository locally for development:
{
"mcpServers": {
"hecras": {
"command": "uv",
"args": ["run", "--directory", "C:\\path\\to\\ras-commander-mcp-main", "ras-commander-mcp"],
"env": {
"HECRAS_VERSION": "6.6"
}
}
}
}
MCP Settings Configuration
In future versions, the MCP server will be able to execute HEC-RAS runs, so the MCP server has settings for the HEC-RAS path, and uses values for HEC-RAS version 6.6 by default. These settings are not yet useful, but will become useful in future versions. To use a different version:
-
Set HEC-RAS Version (if you have a different version installed):
{ "mcpServers": { "hecras": { "command": "uvx", "args": [ "--from", "ras-commander-mcp@git+https://github.com/gpt-cmdr/ras-commander-mcp.git", "ras-commander-mcp" ], "env": { "HECRAS_VERSION": "6.5" } } } } -
Set HEC-RAS Path (if HEC-RAS is installed in a non-standard location):
{ "mcpServers": { "hecras": { "command": "uvx", "args": [ "--from", "ras-commander-mcp@git+https://github.com/gpt-cmdr/ras-commander-mcp.git", "ras-commander-mcp" ], "env": { "HECRAS_PATH": "C:\\Program Files\\HEC\\HEC-RAS\\6.5\\HEC-RAS.exe" } } } }
Testing Configuration
Before adding to Claude Desktop, test your configuration:
# For local development:
cd path/to/ras-commander-mcp-main
set HECRAS_VERSION=6.6
uv run ras-commander-mcp
# Should start successfully showing:
# Starting HEC-RAS MCP Server...
# RAS Commander MCP by CLB Engineering Corporation
Usage
Available Tools
All tools provided by this MCP server leverage the ras-commander Python library for advanced HEC-RAS automation capabilities.
-
hecras_project_summary: Get comprehensive or selective project information
- Parameters:
project_path(required): Full path to HEC-RAS project foldershow_rasprj(optional): Show project file contents (default: true)show_plan_df(optional): Show plan files and metadata (default: true)show_geom_df(optional): Show geometry files (default: true)show_flow_df(optional): Show steady flow data (default: true)show_unsteady_df(optional): Show unsteady flow data (default: true)show_boundaries(optional): Show boundary conditions (default: true)show_rasmap(optional): Show RASMapper configuration (default: false)showmore(optional): Show all columns/verbose mode (default: false)
- Parameters:
-
read_plan_description: Read multi-line description from a plan file
- Parameters:
project_path(required): Full path to HEC-RAS project folderplan_number(required): Plan number (e.g., '1', '01', '02')
- Parameters:
-
get_plan_results_summary: Get comprehensive results from a specific plan
- Parameters:
project_path(required): Full path to HEC-RAS project folderplan_number(required): Plan number or full path to plan HDF file
- Parameters:
-
list_profiles: List available steady profiles or unsteady output time steps for a computed plan
- Parameters:
project_path(required): Full path to HEC-RAS project folderplan_number(required): Plan number or full path to plan HDF file
- Parameters:
-
get_plan_summary: Get plan-level result statistics, including max WSEL, peak flow, volume accounting, and runtime timing
- Parameters:
project_path(required): Full path to HEC-RAS project folderplan_number(required): Plan number or full path to plan HDF filemax_rows(optional): Maximum rows to show in each table (default: 100)
- Parameters:
-
get_mesh_results: Get 2D mesh cell results for a plan profile/time step or maximum-over-time profile
- Parameters:
project_path(required): Full path to HEC-RAS project folderplan_number(required): Plan number or full path to plan HDF fileprofile(optional): Output index, datetime fromlist_profiles, ormax(default: "max")mesh_name(optional): 2D flow area name, or blank for all meshesvariables(optional): Comma-separated variables such asWater Surface,Depth,Velocitymax_rows(optional): Maximum rows to show per mesh (default: 100)
- Parameters:
-
get_xsec_results: Get 1D cross-section WSEL, flow, and velocity results for a plan profile/time step
- Parameters:
project_path(required): Full path to HEC-RAS project folderplan_number(required): Plan number or full path to plan HDF fileprofile(optional): Output index, datetime/profile name fromlist_profiles, ormax(default: "max")variables(optional): Comma-separated variables such asWater_Surface,Flow,Velocity_Totalmax_rows(optional): Maximum rows to show (default: 100)
- Parameters:
-
get_compute_messages: Get computation messages and performance metrics
- Parameters:
project_path(required): Full path to HEC-RAS project folderplan_number(required): Plan number or full path to plan HDF file
- Parameters:
-
get_hdf_structure: Explore HDF file structure
- Parameters:
hdf_path(required): Full path to the HDF filegroup_path(optional): Internal HDF path to start exploration from (default: "/")paths_only(optional): Show only paths without details (default: false)
- Parameters:
-
get_projection_info: Get spatial projection information (WKT)
- Parameters:
hdf_path(required): Full path to the HDF file
- search_docs: Search the ras-commander documentation and return the top matching pages with excerpts (read-only docs retrieval; requires network)
- Parameters:
query(required): Search terms (e.g., "cross section results")
- get_doc_page: Retrieve the markdown/text content of a documentation page by path (read-only docs retrieval; requires network)
- Parameters:
path(required): Page path, e.g.reference/dataframe-reference(leading/trailing slashes optional)
- list_geometry_elements: List major geometry elements from geometry HDF files
- Geometry types:
rivers_reaches: 1D river and reach alignment linescross_sections: 1D cross-section cut lines by river, reach, and stationreference_lines: 2D profile-output reference lines, optionally filtered by mesh areabc_lines: 2D flow area boundary condition linesbreaklines: 2D mesh enforcement linesstructures: bridges, culverts, inline structures, and lateral structuresmesh_areas: named 2D flow areas
- Parameters:
project_path(required): Full path to HEC-RAS project foldergeometry_number(optional): Geometry number such as1,01, org01; full geometry HDF paths are also accepted. Defaults to all geometry HDF files found in the project.element_type(optional): One of the geometry types above, common aliases such asxsorboundary_lines, orall(default)mesh_name(optional): Filter 2D reference lines to a specific mesh areashowmore(optional): Show more attributes while still omitting raw geometry/detail columns (default: false)
Docs tools (
search_docs,get_doc_page) are read-only documentation retrieval and stay within the MCP's stated scope (introspection / QAQC / review — never HEC-RAS execution). They live-fetch from the docs site and therefore require network access. The base URL defaults tohttps://rascommander.infoand is overridable via theRASCOMMANDER_DOCS_URLenvironment variable.
Example Usage in Claude
Once configured, you can ask Claude:
- "Query the HEC-RAS project at C:/Projects/MyRiverModel"
- "Show me the plans in the Muncie test project"
- "Get the results summary for plan '01' in my project"
- "List the output profiles for plan '01'"
- "Show max 2D mesh WSEL, depth, and velocity for plan '01'"
- "Get cross-section WSEL and peak flow results for plan '01'"
- "Show me the compute messages for plan '1'"
- "Explore the HDF structure of my results file"
- "Get the projection info from my terrain HDF"
- "List all geometry elements for the model"
- "List cross sections for geometry g01"
- "Show 2D boundary lines and mesh areas"
- "Search the ras-commander docs for cross section results"
- "Show me the dataframe-reference documentation page"
Python Library Reference
This MCP server is built on top of the ras-commander Python library, which provides comprehensive programmatic access to HEC-RAS projects. For advanced Python scripting and automation beyond what's available through the MCP interface, refer to the ras-commander documentation.
Testing
Running Tests
From the project directory, run the complete test suite:
# Install dependencies and run all tests
uv sync
uv run python tests/test_server.py # Basic server functionality
uv run python tests/test_all_tools.py # Comprehensive tool validation
uv run python tests/test_single_tool.py # Single tool testing utility
Test outputs are saved to tests/outputs/ as markdown files for review.
Test Data
The testdata/ folder contains complete HEC-RAS projects for testing:
Muncie/: Complete project with terrain, results, and boundary conditionsBeaverLake/: Additional test project
Troubleshooting
- ImportError for ras-commander: Ensure HEC-RAS is properly installed
- Project not found: Verify the project path exists and contains .prj files
- Version errors: Check that the specified HEC-RAS version matches your installation
- MCP connection issues: Verify Claude Desktop configuration and restart Claude
Development
To modify or extend the server:
- Clone the repository
- Make changes to
server.py - Test with:
uv run python tests/test_all_tools.py - Update Claude Desktop configuration if needed
License
This project is licensed under the MIT License - see the LICENSE file for details.
Trademarks
See TRADEMARKS.md for trademark information and compliance policies.
About
RAS Commander MCP is developed and maintained by CLB Engineering Corporation as part of our commitment to advancing H&H automation through open-source tools.
For professional H&H automation services and custom solutions, contact us at info@clbengineering.com
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.