mcpField
D
dfxm-geo-mcp

dfxm-geo-mcp

Local
github.com/borgi-sScience & Education0Updated 17w ago
Repository

About dfxm-geo-mcp

dfxm-geo-mcp lets AI clients configure and run the dfxm-geo dark-field X-ray microscopy forward model. It offers reflection and visibility analysis plus browser-ready simulation previews.

Key features

  1. 01Validate and scaffold configurations
  2. 02Find reachable reflections
  3. 03Rank reflection visibility
  4. 04Render forward-model previews
  5. 05Run rocking-curve simulations
  6. 06Launch Monte Carlo jobs

Use cases

  1. 01Check a model configuration before simulation
  2. 02Identify reachable reflections for an experiment
  3. 03Preview simulated images and rocking curves

MCP support

An MCP server that exposes dfxm-geo configuration, reflection analysis, simulation, and job-management operations to AI clients.

Configuration

Local server config is the same for every client. Copy it into your client’s MCP config file.

mcp.json · Universal config
{
  "mcpServers": {
    "borgi-s-dfxm-geo": {
      "command": "uvx",
      "args": [
        "dfxm-geo-mcp"
      ]
    }
  }
}
Run commanduvx dfxm-geo-mcp

Local servers need a runtime on your machine (e.g. Node.js 18+). Config file locations differ by client; see your client’s docs.

Packages
pypidfxm-geo-mcp0.1.0

Tools

No tool list yet

Local servers expose tools only once running on your machine, and the maintainer has not provided a tool list.

Details

From the project README

dfxm-geo-mcp

An MCP server that lets an AI client drive the dfxm-geo dark-field X-ray microscopy forward model: validate and scaffold configs, enumerate reachable reflections, and render preview-scale simulations.

Claude Desktop rendering a DFXM image from a plain-English request

Run

uvx dfxm-geo-mcp

Claude Desktop (claude_desktop_config.json):

{"mcpServers": {"dfxm-geo": {"command": "uvx", "args": ["dfxm-geo-mcp"]}}}

Or: pip install dfxm-geo-mcp then dfxm-geo-mcp.

First run pulls a heavy scientific stack (numba/scipy) and warms a JIT cache (~10 s once). No fake demo mode — the sims are real.

Tools

validate_config · find_reflections · predict_visibility (g·b reflection-visibility ranking; structured scores plus a self-contained .html) · scaffold_config · run_forward (analytic preview; saves a PNG and a self-contained .html) · run_rocking (interactive φ-rocking-curve viewer: a single self-contained .html with a frame scrubber + live rocking-curve plot) · start_bootstrap / get_job_status / get_job_result (MC fidelity).

Both run_forward and run_rocking write self-contained HTML (image(s) embedded, all CSS/JS inline, no external origins) that opens full-size in any browser and is surfaced by file-showing clients.

Architecture

A protocol-agnostic ops layer wrapping dfxm-geo, under a thin FastMCP adapter. See docs/superpowers/specs/.

Roadmap (v2)

run_identify · remote HTTP transport · .mcpb bundle.

Comments

Markdown supported

New to MCP?

From what the protocol is to installing your first server, these answers cover most first questions.

An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.