MDEngine

MDEngine

*by ForceField Silicon*

A molecular-dynamics workbench with a LAMMPS MCP server for Claude and other agents: a native macOS Metal trajectory viewer, a command-line tool, and an MCP server that inspects trajectories and runs LAMMPS simulations as detached background jobs. The hosted tier runs the same jobs on a GPU: Claude sends a deck to the molecular dynamics MCP server at api.forcefieldsilicon.com and gets the results back, with nothing to install.

Start here

• Claude Code: claude mcp add --transport http mdengine-cloud https://api.forcefieldsilicon.com/mcp
• claude.ai: add a custom connector with the URL https://api.forcefieldsilicon.com/mcp
• Claude Desktop: install the .mcpb bundle from the GitHub release
• Your key comes with a credit pack at https://forcefieldsilicon.com/mdengine
• Campaign guides for mineral processing and protein work: https://github.com/forcefieldsilicon/mdengine-plugins

Status: 0.7.1 released. Signed and notarized macOS app, command-line tool, Claude Desktop

bundle, and the hosted GPU tier. Release notes: https://github.com/forcefieldsilicon/mdengine/releases

Parts

Build

Requires macOS 14+ and Xcode command-line tools. LAMMPS (brew install lammps) is needed only for run / submit_lammps.

swift build -c release            # all three products

./scripts/make_app.sh # assemble + install /Applications/MDEngine.app

Bundled examples

Trajectory readers are safe on in-flight dumps — a file a running simulation is still writing parses to its complete frames, so you can inspect a run mid-flight. Rows with non-finite (NaN/inf) coordinates are dropped.

Both decks are ready inputs for mdengine run and submit_lammps. Bare force-field names resolve automatically: if $LAMMPS_POTENTIALS is unset, MDEngine derives it from the LAMMPS install.

MCP server

mdengine-mcp is a dependency-free MCP stdio server, listed in the MCP Registry as com.forcefieldsilicon/mdengine (server.json in this repo). Each release ships a signed macOS MCP Bundle (mdengine-mcp- -macos-arm64.mcpb): double-click it to install in Claude Desktop, or unpack it (mcpb unpack) for any other client. Built from source, register with Claude Code:

claude mcp add mdengine /path/to/.build/release/mdengine-mcp

or in Claude Desktop's claude_desktop_config.json:

{ "mcpServers": { "mdengine": { "command": "/path/to/.build/release/mdengine-mcp" } } }

Hosted GPU tier over MCP (no install)

The same jobs are reachable from any MCP client that speaks HTTP — Claude Code, Claude.ai custom connectors, Cursor, Goose — via the hosted endpoint's Streamable HTTP server:

claude mcp add --transport http mdengine-cloud https://api.forcefieldsilicon.com/mcp

Sign in when the client asks (OAuth 2.1: a consent page where you paste your API key once; the client keeps a token, the key stays with you). Scripted clients may instead send the key directly as --header "Authorization: Bearer mde_YOUR_KEY". The key comes with a prepaid credit pack (forcefieldsilicon.com/mdengine). initialize and tools/list work without signing in; tool calls without a credential return 401 with the OAuth Windows and Linux. The hosted tier is the supported path on both, and it is the full paid product: every tool below works from Claude Code on Windows exactly as on a Mac. In PowerShell, register at user scope so the server follows you into every folder, then sign in once:

claude mcp add --scope user --transport http mdengine-cloud https://api.forcefieldsilicon.com/mcp

Start claude, type /mcp, pick mdengine-cloud, choose Authenticate, and paste the key on the browser consent page (never into the chat). Decks written on Windows (CRLF line endings) are accepted as-is. The local viewer, renderer and CPU job runner are macOS-only today; a Windows/Linux local build is planned and demand decides its order, so say so if you need it.

discovery pointer, which is what makes clients offer the sign-in. Tools: account, capabilities, preflight_deck, submit_job (deck inline, ≤ 8 MB), create_job + start_job (big decks via presigned PUT), job_status, job_log, job_results, list_jobs, delete_results, cancel_job (12 tools). Discovery card: https://api.forcefieldsilicon.com/.well-known/mcp/server-card.json. The hosted service (endpoint, billing, GPU launcher) is operated by ForceField Silicon and its source is not part of this repository; the client side of it (HostedClient, CLI run --gpu, MCP host=cloud) is here under the same license as everything else.

Preflight before spend (GJOB-118). GET /v1/capabilities publishes the hosted image's LAMMPS version, packages and every style with a gpu flag: true = KOKKOS-accelerated, false = the style exists but runs on the pod's CPU cores at the GPU rate. Every client checks a deck against it first — mdengine capabilities deck.in, the app's Run Accelerated, submit_lammps host=cloud, the hosted submit_job — refusing a deck whose styles the image lacks (LAMMPS would exit at startup) and warning when the pair style has no /kk variant (the run would not use the GPU). --force / force=true overrides; the server only advises (preflight in the start response), it never blocks.

Troubleshooting the hosted connector:

Jobs run in the deck's own directory (relative read_data paths work) and launch with -sf omp -pk omp N so the OPENMP package is actually engaged; $LAMMPS_POTENTIALS is derived from the LAMMPS install when unset.

Remote hosts (run on your own Linux / GPU box)

The job runner can execute on another machine with the same tool contract: declare hosts in ~/.mdengine/hosts.json and pass host to submit_lammps (or set a default). The deck's directory is rsynced up (trajectories, checkpoints and logs excluded), LAMMPS starts under nohup with its exit code recorded remotely, and job_status / job_log / job_files / cancel_job work unchanged; fetch_job brings results back for trajectory_info, z_profile and the renderers.

{

"default": "gpu1", "hosts": { "gpu1": { "ssh": "me@gpu1.example.net", "workdir": "~/mdengine-jobs", "lmp": "/usr/local/bin/lmp", "threads": 8, "launch": "{lmp} -in {input} -k on g 1 -sf kk -pk kokkos newton on neigh half -log {log}" } } }

ssh is anything ssh accepts (key-based, non-interactive); lmp must be an absolute path (login PATH is not available over ssh); launch is optional — the default is the OpenMP form, and a GPU host is simply one whose template carries the KOKKOS flags. A remote deck must be self-contained within its own directory. The same trust model as local runs applies, on the remote machine.

Platform & limits

macOS 14+ (Apple Silicon or Intel). Trajectories are loaded whole into memory — files over 2 GB are refused with guidance to decimate or split first.

Security note

Running a LAMMPS input executes whatever the deck says. LAMMPS decks are programs, not data — they can invoke arbitrary shell commands (LAMMPS has a literal shell command). Treat a deck from someone else exactly like a shell script: read it before running it.

This applies doubly to the MCP server: an AI agent connected to `mdengine-mcp` can submit decks, and a submitted deck runs with your user's full privileges on this machine. That is the same trust model as any local dev tool (an agent that can run make can run anything), but be deliberate about which decks — and which agents — you hand to the job runner. Sandboxed execution (containers, no network, resource caps) is how a future hosted tier makes running untrusted decks safe; the local server does not sandbox.

Privacy

The local tools collect nothing; see PRIVACY.md, which also covers the hosted GPU tier.

License

Business Source License 1.1 (BUSL-1.1), © 2026 Gitinama Inc. (ForceField Silicon). See LICENSE.

In short: you may read, copy, modify and redistribute MDEngine, and you may use it in production for your own research, engineering or commercial work, on your own machines or on machines you rent. The one thing the li