MDEngine
https://api.forcefieldsilicon.com/mcp ↗
Description
MD workbench for agents: 13 analysis tools, renders, hosted GPU runs; LAMMPS + OpenMM.
What our measurements say
MDEngine is a remote MCP server published at api.forcefieldsilicon.com. It has been probed 6 times since 9/12/2026. It answered in 6 of them (100.0%), a near-uninterrupted record. Median response time is 191 ms, placing it among the faster endpoints. It offers a narrow, focused set of 12 tools. On the protocol side it still runs 2025-11-25 and has not moved to the newer spec.
Agent readability score
Can an LLM agent pick the right tool here — names, descriptions and parameter clarity are assessed.
Agent readability score
Strengths
- Most tool parameters are well-defined with required/optional flags and types.
- Job lifecycle tools (submit_job, create_job, start_job) are logically separated.
Weaknesses
- Tool names like 'account' are too generic and don't clearly indicate purpose.
- Descriptions often focus on implementation details rather than usage scenarios.
- Some parameters (e.g., 'runner' in submit_job) lack explicit value examples.
account— Name is too generic, doesn't indicate balance or API key management purposecapabilities— Description doesn't clarify when to use it for runner selectionpreflight_deck— Parameter 'files' lacks format specification in descriptionsubmit_job— Multiple parameters (gpu, runner) lack explicit value examples
Verdict
Event narrative
Risk: low
Advice
Schema compliance
tools/list structure, inputSchema validity, and a functional smoke test — the components of the 0-100 score.
- tools/list schema valid
- inputSchema is valid JSON Schema
- Smoke test not run this round
- Cold-start: cold 4ms → warm 3ms (penalty 1ms)
Tools
Tools the server advertised in the latest measurement — measured, not catalog-claimed.
accountBalance in USD, how jobs are priced (`pricing.mode` job = the deck's own work in atom-steps at per-class prices, capped at wall_limit_s x rate; metered = per second of pod time), the rate table, and the key id of the API key in use.
submit_jobOne call: create a hosted GPU job, upload the deck given INLINE as {relative_path: text}, and queue it. Total inline size <= 8 MB; for larger decks use create_job, PUT the tarball to upload_url, then start_job. Billing starts at the first heartbeat (state running) and stops at done/failed/cancelled. The deck is executed as a program on an isolated GPU pod.
inputstringrequiredlabelstringgpustringwall_limit_sintegerestimate_sintegerrunnerstringlaunchstringfilesobjectrequired
create_jobStep 1 of the two-step path for big decks: validates the spec, reserves a job id, returns a presigned upload_url. PUT the deck as a .tar.gz (<= 2 GB, relative paths, input at `input`) to upload_url, then call start_job.
inputstringrequiredlabelstringgpustringwall_limit_sintegerestimate_sintegerrunnerstringlaunchstring
start_jobStep 2: queue a job whose deck tarball has been uploaded. A GPU pod is launched; billing starts when it reports running.
idstringrequired
job_statusState (created|uploaded|queued|launching|running|uploading|done|failed|cancelled), GPU, rate, billed seconds, cost so far, exit code, error, last thermo lines.
idstringrequired
job_logThe last <= 20 thermo/log lines the running pod reported (30 s heartbeat). Full log.lammps is in the results tarball.
idstringrequired
job_resultsFor a done/failed job: a presigned download_url (valid ~7 days) for the results tarball (work/, log.lammps, exitcode). Results are deleted 30 days after the run.
idstringrequired
list_jobsJobs of this API key, newest first. Compact rows (id, state, label, created, finished, gpu, cost_usd, error) unless full=true.
limitintegerfullboolean
delete_resultsFor a finished job: delete its deck and results tarballs immediately instead of at the automatic 30-day purge. Metadata and billing records are kept; results can no longer be downloaded.
idstringrequired
capabilitiesCapability manifest of the hosted runners: LAMMPS version, installed packages, and every style by category with gpu=true (KOKKOS-accelerated) or gpu=false (exists, but runs on the pod's CPU cores at the GPU rate). Default = compact summary; runner=lammps&full=true returns the whole style table. Use preflight_deck to check a specific deck.
runnerstringfullboolean
preflight_deckDry run of the check submit_job performs: which styles the deck asks for are MISSING on every hosted LAMMPS image (the run would exit at startup), which are CPU-only, whether its pair style will use the GPU at all, and which image (`runner`) the job will be routed to — decks needing packages beyond the fast default image run on lammps-full automatically. Nothing is created or billed.
inputstringrequiredfilesobjectrequired
cancel_jobCancel a job that is not finished. A running job is billed up to the cancel time; its pod is terminated.
idstringrequired
Availability history
Event log
Derived by comparing consecutive probes — changes in era, protocol version, build and reachability.
- Added to watch — era: Legacy
Get your badge
Add this badge to your README — it updates automatically as measurements change.
[](https://mcpmetrics.io/servers/com-forcefieldsilicon-mdengine)<a href="https://mcpmetrics.io/servers/com-forcefieldsilicon-mdengine"><img src="https://mcpmetrics.io/badge/com.forcefieldsilicon/mdengine/era.svg" alt="mcpmetrics"></a>You are seeing the last 7 days. Sign up for the full history. Which check failed and why is in the dashboard.
Sign up free to seeSimilar MCP servers
The catalog entries whose name and description are closest to this one, found with the same index the search box uses.
- IA-QA — 130+ QA & Dev Tools for AI Agentswww.ia-qa.com
130+ QA & dev tools for AI agents: prompt injection, RAG testing, VLM eval, guardrails. Free.
- Tools for Agentsapi.toolsforagents.tools
Decision Layer for AI Agents — 58+ tools, Advisor, MCP. Free key: POST /v1/register {}.
- dev.workers.dolerite.base-onchain-tools/base-onchain-toolsbase-onchain-tools.dolerite.workers.dev
Base mainnet analysis: contract audits, transfer simulation, token pricing. Paid via x402.
- BubblyPhone Agentsagents.bubblyphone.com
Make AI phone calls from any AI assistant. BubblyPhone's MCP server lets you manage AI-powered phone agents — book reservations, schedule appointments, set up customer support lines, or have your AI agent call anyone on your behalf. ## Features - **Make AI phone calls** — Initiate outbound calls with AI agents (Gemini, GPT) that handle the conversation - **Manage phone numbers** — Search, purchase, and configure numbers in 30+ countries - **Configure AI agents** — Set system prompts, choose AI models, select voices - **Monitor calls** — View call events, transcripts, and recordings - **Track billing** — Check balance, usage, and transaction history - **Mid-call control** — Inject context, transfer calls, or hang up ## 20 Tools `make_call` · `list_calls` · `get_call` · `hangup_call` · `transfer_call` · `inject_context` · `get_call_transcript` · `get_call_events` · `search_phone_numbers` · `list_phone_numbers` · `buy_phone_number` · `get_phone_number` · `update_phone_number` · `get_b
- MCP LatAm Toolsmcp-latam-tools-production.up.railway.app
A collection of validation and utility tools for Latin American markets. Covers Brazil (CPF, CNPJ, PIX key validation, national holidays), Mexico (RFC validation, national holidays), and Chile (RUT validation, national holidays). Designed for agents and applications handling tax forms, invoices, payments, e-commerce, and compliance workflows across Latin America.
- AIGEN — Open Bounty Protocol for AI Agentscryptogenesis.duckdns.org
Comments
Sign in to write a comment
No comments yet. Be the first.