Serveur MCP MDEngine
Commit 29 oct. 2026Permet à un client IA d'inspecter des trajectoires de dynamique moléculaire, de générer des vidéos de simulation et d'exécuter des simulations LAMMPS sous forme de tâches de fond détachées.
Aussi connu sous forcefieldsilicon/mdengine
- Étoiles
- 0
- Outils
- —
- Téléchargements, 30 jours
- —
- Dernière version
- 0.7.1
- Dernier commit
- 29 oct. 2026
- Licence
- —
Qu'est-ce que MDEngine MCP ?
MDEngine est un serveur MCP. Permet à un client IA d'inspecter des trajectoires de dynamique moléculaire, de générer des vidéos de simulation et d'exécuter des simulations LAMMPS sous forme de tâches de fond détachées.
Résumé généré à partir du README du projet
À propos du MCP server MDEngine
MDEngine fournit un environnement de travail spécialisé pour la dynamique moléculaire, permettant aux agents IA d'inspecter des trajectoires et de gérer des simulations. Les développeurs peuvent utiliser le serveur local macOS via stdio pour exécuter des simulations LAMMPS en tant que tâches de fond détachées qui se poursuivent même si la machine se met en veille. Le système inclut des outils pour analyser le nombre d'atomes, les bounding boxes et les plages de charges, ainsi qu'une analyse de profondeur spécialisée pour les processus d'oxydation ou de dépôt.
Pour les utilisateurs sur Windows ou Linux, ou ceux ne disposant pas de matériel local, une offre GPU hébergée est disponible via un transport HTTP. Cela permet à des clients comme Claude Code ou Cursor d'envoyer des jeux de simulation vers un endpoint distant et de recevoir les résultats sans aucune installation locale. Les utilisateurs peuvent gérer ces tâches distantes à l'aide d'outils pour vérifier le statut, consulter les logs ou annuler des exécutions. L'accès à l'offre hébergée nécessite une clé API associée à un pack de crédits, laquelle peut être gérée via un processus de consentement OAuth 2.1.
Ce que vous pouvez faire avec
- 1Analyser la profondeur de simulationUn agent IA peut calculer la profondeur de pénétration d'une sonde lors d'un processus d'oxydation.
- 2Générer des visuels de simulationUn développeur peut demander une vidéo MP4 ou un GIF animé d'une trajectoire moléculaire.
- 3Exécuter des simulations GPU à distanceUn agent peut soumettre un deck LAMMPS à un GPU basé sur le cloud pour effectuer des calculs intensifs.
- 4Inspecter des données en cours de générationUn chercheur peut examiner un fichier de trajectoire pendant qu'une simulation est encore en train d'y écrire activement.
Extrait du README du projet. Chaque cas d'utilisation est vérifié par rapport à une citation du document.
Comment installer MDEngine MCP
Formats vérifiés le 10 nov. 2026Ajouter à CursorAjouter à GooseInstaller MDEngine MCP dans Claude Code
$ claude mcp add --transport http mdengine https://api.forcefieldsilicon.com/mcpInstaller MDEngine MCP dans Cursor
~/.cursor/mcp.json
{
"mcpServers": {
"mdengine": {
"url": "https://api.forcefieldsilicon.com/mcp"
}
}
}Installer MDEngine MCP dans VS Code
.vscode/mcp.json
{
"servers": {
"mdengine": {
"type": "http",
"url": "https://api.forcefieldsilicon.com/mcp"
}
}
}Installer MDEngine MCP dans VS Code CLI
$ code --add-mcp '{"name":"mdengine","type":"http","url":"https://api.forcefieldsilicon.com/mcp"}'Installer MDEngine MCP dans Codex
$ codex mcp add mdengine --url https://api.forcefieldsilicon.com/mcpInstaller MDEngine MCP dans Codex config.toml
~/.codex/config.toml
[mcp_servers.mdengine] url = "https://api.forcefieldsilicon.com/mcp"
Installer MDEngine MCP dans Gemini CLI
$ gemini mcp add -s user --transport http mdengine https://api.forcefieldsilicon.com/mcpInstaller MDEngine MCP dans Gemini settings.json
~/.gemini/settings.json
{
"mcpServers": {
"mdengine": {
"url": "https://api.forcefieldsilicon.com/mcp",
"type": "http"
}
}
}Vérifications automatisées
Exécuté le 14 nov. 2026Rapport complet →Pas encore vérifié. Un serveur nécessite un score de 80, aucune faille sérieuse et suffisamment de données pour que nous puissions effectuer les vérifications. Fonctionnement des vérifications · Ce qu'il faut corriger pour gagner des points
- Dépôt source lié et accessibleforcefieldsilicon/mdengine is public and reachablegithub.com/forcefieldsilicon/mdengine ↗Réussi
- Le nom du package correspond au dépôt sourceNon vérifié
- Provenance du build (attestation Sigstore)Non vérifié
- Licence open source (OSI)No licenseFaille
- Commit ou version publiée au cours de l'année écouléeLast commit or release 2026-10-29github.com/forcefieldsilicon/mdengine ↗Réussi
- OpenSSF ScorecardNon vérifié
- Nom du package non proche d'un package populaireNon vérifié
- Aucun avis critique dans le package ou ses dépendancesNon vérifié
- Aucune instruction cachée destinée au modèle0 tool descriptions and the README read; no override phrases, credential paths or hidden charactersRéussi
- Les outils destructifs sont étiquetésNon vérifié
- Le point de terminaison distant répond à la poignée de main MCPinitialize answered with protocol 2025-06-18 in 471 msRéussi
README
Voir sur GitHubMDEngine
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
.mcpbbundle 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
| Product | What it is |
|---|---|
| MDEngine.app | Metal viewer — orbit / pan / zoom camera, trajectory timeline with 20%/5% tick marks, CPK element colours, XYZ / extended-XYZ, live settings |
| mdengine-cli | info / export / decimate / run on trajectories and decks |
| mdengine-mcp | MCP stdio server — trajectory tools plus a detached LAMMPS job runner (submit_lammps / job_status / job_log / list_jobs / cancel_job) |
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
| Example | What it shows |
|---|---|
lj_melt.xyz / lj_melt.in |
Lennard-Jones argon melt — the minimal smoke test |
fe_oxidation.xyz / fe_oxidation.in |
ReaxFF iron oxidation: a bcc Fe slab meeting O₂ gas, using ffield.reax.Fe_O_C_H from the LAMMPS distribution |
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-<version>-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" } } }
| Tool | Does |
|---|---|
trajectory_info |
Frames, atom counts, per-atom fields, elements, bbox, charge range |
z_profile |
Deposition/oxidation depth analysis: substrate surface plane, probe penetration depths (min/mean/max), at-surface & in-flight counts, bound-probe charge, z histogram |
render_video |
Trajectory → MP4 (H.264) or animated GIF via the Metal renderer: camera angles, stride, cinematic orbit, baked scale-bar/frame annotations, per-element colors/sizes, style: "contrast" auto-visibility |
render_image |
One frame → PNG with the same camera/style options — lets an agent see a simulation state |
export_frame |
One frame → XYZ; charges: true → extended-XYZ with the q column |
decimate |
Keep every Nth frame (final frame always kept) |
submit_lammps |
Detached LAMMPS job: survives the server exiting and machine display-sleep (caffeinate), exit code recorded unattended |
job_status / job_log |
State + live thermo tail / raw log tail |
job_files |
Locate a finished job's outputs (run dir + bookkeeping dir) |
list_jobs / cancel_job |
Registry under ~/.mdengine/jobs/ / SIGTERM a run (state becomes cancelled) |
list_hosts |
Execution hosts from ~/.mdengine/hosts.json and which is the default |
fetch_job |
Pull a remote job's run directory (dumps, data) + logs into the local job dir under results/ |
run_lammps |
Synchronous run for short tests only |
buy_credits |
Opens the credit page to top up the hosted GPU key |
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:
| Symptom | Cause / fix |
|---|---|
| Client says authentication required / 401 | Use the client's sign-in (OAuth) and paste your mde_… key on the consent page, or add the header Authorization: Bearer mde_…. Keys are issued at purchase and shown once. |
| Consent page says the key was not recognised | Keys are mde_ + 32 hex characters; a revoked key no longer works. mdengine account prints the balance for a saved key. |
| "insufficient balance" | Top up at forcefieldsilicon.com/mdengine; submissions need credit for at least 15 min at the GPU rate. |
| "gpu_runners_open_soon" (503) | Runners are temporarily closed; credits are safe. |
| HTTP 429 | Too many keyless requests from one IP; add the key or slow down. |
job_results says results unavailable |
Files were deleted by delete_results, the run produced none, or the 30-day purge ran. |
Job failed with runner_error / lammps_error |
The deck itself failed; job_results (if present) or job_log holds the LAMMPS error text. |
Job failed with pod_lost / no_capacity / launch_timeout / gpu_unavailable |
Infrastructure (pod died, no GPU stock, a pod that never came up, or the host handed the pod no CUDA device); never billed, relaunched once automatically. Resubmit if it still fails. |
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 license withholds is offering MDEngine itself, or a service whose value comes mainly from it, to third parties as a hosted, managed or embedded service. Each version converts to the Mozilla Public License 2.0 four years after its first public release. Versions 0.7.0 and earlier were released under MIT and stay MIT. Other arrangements: arvand@gitinama.tech.
Questions sur MDEngine MCP
Comment ajouter MDEngine MCP à Claude Code ?
Exécutez claude mcp add --transport http mdengine https://api.forcefieldsilicon.com/mcp dans votre terminal. Ajoutez --scope user pour le rendre disponible dans tous les projets.
Comment ajouter MDEngine MCP à Cursor ?
Ouvrez ~/.cursor/mcp.json et ajoutez le serveur sous mcpServers, avec la configuration Cursor indiquée dans la section d'installation ci-dessus.
Puis-je exécuter des simulations sur un GPU sans rien installer ?
Oui, l'offre hébergée vous permet d'envoyer des jeux de simulation vers l'endpoint GPU hébergé et de recevoir les résultats via un client MCP.
Est-il sûr d'inspecter une simulation pendant qu'elle est encore en cours d'exécution ?
Oui, les lecteurs de trajectoires peuvent analyser en toute sécurité des dumps en cours de génération, ce qui vous permet d'inspecter une exécution en plein milieu en lisant les frames complètes.
Quels sont les risques de sécurité liés à l'utilisation du lanceur de tâches LAMMPS ?
Les decks LAMMPS sont des programmes capables d'invoquer des commandes shell arbitraires ; un agent IA connecté au serveur peut exécuter des decks avec l'intégralité des privilèges de votre utilisateur.
Comment s'authentifier avec l'offre GPU hébergée ?
Vous pouvez utiliser OAuth 2.1 via une page de consentement où vous collez votre clé API, ou passer la clé directement en tant que Bearer token dans l'en-tête d'autorisation.