Run ED staffing simulations inside your Claude chat.
The QSimHealth MCP server lets Claude, and any AI app that speaks MCP, run healthcare staffing simulations in the conversation. Describe your ED, walk-in clinic or appointment office. Claude runs the simulation and reads back the numbers. Free demo, no sign-up, no API key.
https://qsimhealth.com/mcp/v1
Paste it into Claude, Cursor or any MCP client. No key needed.
Ask Claude a staffing question and get real numbers back.
ED, walk-in or urgent care clinic, and appointment office
Each one has its own staffing problems. ED triage breaks first-come first-served, walk-ins need the right MD/PA mix, appointment offices fight no-shows. Claude picks the right teaching tool.
Open the scenario in the QSimHealth app
Every result includes a link that opens the scenario in the QSimHealth app with your inputs filled in and the chart drawn. New users sign up in about 30 seconds for a 10-day free trial; signed-in users go straight to the simulator.
Numbers from the engine, not the model's memory
Every number comes from a discrete-event simulation on the ChiAha clinic engine. The system prompt, tool descriptions and a provenance note on each result tell Claude to quote those numbers exactly instead of estimating.
AI assistants can run and experiment with staffing models on our simulation engine.
Anyone can get a simulation that runs. The hard part is one that matches your facility. Instead of asking Claude to write patient-flow code from scratch, let it drive the ChiAha clinic engine, a discrete-event engine from the same family as QueueSim’s, which QueueSim checks against closed-form Erlang C results with its public compare_analytical_vs_simulated tool. Ask in plain English; Claude runs ED demo scenarios and finds the smallest MD count that meets a wait target, and the numbers come from the simulation.
Building your own model with an assistant? Validating it takes your facility's measured arrivals by hour and treatment times, a check against the waits you actually recorded over weeks, hour by hour, and a test on a period the model has never seen. What validation takes.
ChiAha’s ReliaSim® was the first commercial production simulation software that AI assistants such as Claude and ChatGPT can use to build models and run experiments on its validated engine. QSimHealth’s MCP server comes from the same team.
Claude is told to quote the engine’s numbers exactly, show the inputs behind every result, and say so when a question is outside what the tools cover. Demo results describe a sample facility, not yours, so they are not staffing advice for a real ED or clinic. AI info.
Use is governed by the ChiAha Software License (/license), including its restrictions on reverse engineering, AI training on outputs and scraping.
Three ways in. Pick the app you already use.
Easiest
Settings → Connectors (or Custom Integrations). Add an MCP server:
https://qsimhealth.com/mcp/v1
No auth. Open a chat and try one of the prompts below.
Config file
Edit claude_desktop_config.json:
{
"mcpServers": {
"qsimhealth": {
"command": "npx",
"args": ["-y", "mcp-remote",
"https://qsimhealth.com/mcp/v1"]
}
}
}
Uses the mcp-remote bridge. Restart Desktop.
Custom MCP
Settings → Features → Model Context Protocol → Add Custom MCP. Use the same URL:
https://qsimhealth.com/mcp/v1
The same steps work for Windsurf and other MCP clients.
What Claude can call once it's installed.
Runs a 7-day demo with constant arrivals and constant MD staffing. Returns hourly results, a summary and a link that opens the scenario in the app. The workhorse.
The inverse problem: given arrivals per hour and a target average wait, returns the smallest MD count that meets it. Counts that would overload are skipped and named in the response.
Four facility types: ED, urgent care, walk-in clinic and appointment office.
Detail for each type: typical arrival pattern, MD/PA mix, common pain points, and what the demo covers compared with a custom model.
Triage breaking first-come first-served, patients who leave without being seen, why peak hours dominate, and what plain M/M/c misses.
Clinics that close for the night, MD/PA mix as the lever, short patient patience and peak-hour coverage.
No-show variation, buffer time as the trade-off, double-booking, and treatment time by visit type.
Demo scope: MD-only staffing and a flat 24-hour arrival rate. For mixed staffing (MD, PA, NP and locum), hourly arrival patterns from your facility, acuity mix, patients who leave, and shift schedules, start a 10-day free trial of the full app, or ask us about a custom facility model.
Once it's installed, ask Claude any of these.
Click a prompt to copy it.
-
Copy
“I run a 24-hour ED with about 8 patients an hour. With 3 MDs and 15-minute average treatment, what's the average wait?”
Calls
simulate_ed_demo. Expect a wait around 3 minutes at about 67% utilization, plus a link to open the scenario in the app. -
Copy
“How many MDs do I need to keep average wait under 15 minutes at 12 patients an hour?”
Calls
recommend_md_count. Returns the smallest MD count that meets the target and names the counts that would overload. -
Copy
“Why does my ED wait get so much worse in the evening peak when arrivals are only 20% higher?”
Calls
explain_ed_queueing. Covers why wait grows much faster than utilization. -
Copy
“My appointment clinic has 12% no-shows and we double-book. How does that change scheduling?”
Calls
explain_appointment_office. Covers buffer time and how it interacts with no-shows.
Use the full app for your facility's real data.
The MCP demo runs flat scenarios. That's good for intuition, not for production decisions. Real ED staffing work needs your hourly arrivals, your MD and PA shifts, treatment times by acuity and how many patients leave without being seen. The QSimHealth app takes those inputs, and ChiAha builds custom facility models on them.
Our team has built simulation software since 1990.