No SQL. No Templates. Just Ask.
Every question gets its own query, its own analysis, and its own visualization, built on the fly. No SQL, no rigid templates, no waiting for someone else to pull it.
Between you and the answer: SQL, an analyst, or a report nobody built yet
You have a question. “How do our readmission rates compare to similar hospitals nearby?” Or “is this locum physician disciplined in any state?” Or “which drugs on our formulary had FDA enforcement actions last year?”
The data to answer these is public. CMS publishes it, FDA publishes it, every state board publishes it. But between you and that data there's a wall: CSV files with cryptic column names, APIs that need a developer, 15 government websites that don't talk to each other. So you submit a ticket. Or hire a consultant. Or just don't get the answer.
And even if you get the data, you still need someone to make sense of it. Most platforms sell you pre-built reports that answer the questions someone else thought you'd ask. If your question doesn't fit the template, you're back to the ticket queue.
Medistill works differently. You ask a question in plain English. It writes the query, pulls the data from 2,000+ structured datasets, and generates a custom report with charts and tables built specifically for your question. Every question gets its own analysis and its own visualization. No SQL, no pre-built templates, no waiting for someone else to build what you need.
The compliance officer screening a provider
Rachel runs compliance at a 320-bed community hospital. When a new physician applies for privileges, she checks OIG LEIE and SAM.gov manually. It takes 15 to 20 minutes per provider, and she knows she's only scratching the surface. State boards, FDA debarment, OFAC sanctions, Open Payments - she doesn't have time to check all of them. She types:
You type
“Screen Dr. James Martinez, NPI 1234567893, against all enforcement and exclusion databases.”
Medistill returns
130+ sources checked in under 30 seconds. The FL board probation would not have surfaced from a standard OIG + SAM.gov check.
Rachel didn't write a query. She didn't open 10 browser tabs. She asked a question and got a structured answer with a risk score. The Florida board probation is something she would have missed entirely, because her hospital is in Texas and she doesn't routinely check Florida's board. Medistill checks all 50 states automatically.
The CFO benchmarking readmissions
David is CFO at a regional medical center in Ohio. CMS penalized his hospital $1.2 million last year under the Hospital Readmissions Reduction Program. He wants to know how his readmission rates compare to peer hospitals, but his analytics team is backlogged for three weeks. He types:
You type
“How do readmission rates at Fairview Regional Medical Center compare to hospitals within 50 miles? Include star ratings, HCAHPS patient satisfaction, and excess readmission ratios for heart failure and pneumonia.”
Medistill returns
| Hospital | Stars | HF ERR | PN ERR | HCAHPS |
|---|---|---|---|---|
| Fairview Regional | 2 | 1.14 | 1.08 | 58% |
| St. Mary's Health | 4 | 0.91 | 0.88 | 72% |
| Ohio Valley General | 3 | 1.02 | 0.95 | 65% |
| Lakeshore Community | 3 | 0.97 | 1.01 | 68% |
| National median | - | 1.00 | 1.00 | 67% |
ERR = Excess Readmission Ratio (above 1.0 means worse than expected). HCAHPS = % of patients rating hospital 9 or 10 overall.
Data sourced from CMS Hospital Compare, Hospital Readmissions Reduction Program, and HCAHPS Survey results. Updated with the latest available reporting period.
David can see immediately that his heart failure excess readmission ratio (1.14) and pneumonia ratio (1.08) are both above the national median, while his nearest competitor, St. Mary's, is below 1.0 on both. His HCAHPS scores are 9 points below the national median.
He didn't build this table. He didn't download a spreadsheet, figure out which CMS dataset has readmission data (it's spread across three), or ask an analyst to design a comparison. The table, the formatting, the peer selection, Medistill generated all of it from his question.
But here's where it gets interesting. David doesn't stop at readmissions. He follows up:
“Show me every provider affiliated with our hospital, grouped by specialty. How does our specialist coverage compare to St. Mary's?”
“Find all cardiologists within 15 miles who are NOT on our roster. Screen them for compliance issues and show their Medicare billing volume, those are our recruitment targets.”
“Now run predictive risk scores on our existing cardiology staff. Who looks clean today but has billing patterns that match providers who got excluded last year?”
One conversation just crossed hospital quality data, a provider-facility crosswalk linking 1.7 million physicians to 273,000 facilities, geographic provider search, compliance screening across 130+ enforcement databases, Medicare billing data, and a predictive risk model trained on 7 million providers. Six separate data dimensions, no SQL, no analyst, no switching between tools.
This is the kind of analysis that used to require a consultant, a data team, and a month on the backlog. David went from a readmission comparison to a recruitment pipeline to a risk assessment in three follow-up questions.
The staffing agency verifying multi-state credentials
Marcus owns a locum tenens staffing agency. He places physicians across Florida, Texas, and California. Every placement requires verifying active licenses and checking for disciplinary actions in each state. That means navigating three different state board websites, each with its own search interface and quirks. He types:
You type
“Dr. Angela Reeves, NPI 1987654320, holds medical licenses in FL, TX, and CA. Verify all three are active and check for any disciplinary actions or board orders.”
Medistill returns
License: ME 148276 | Expires: 2027-01-31
License: R7831 | Expires: 2026-08-31
License: A 94217 | Expires: 2026-12-31
Stipulated settlement, effective 2024-06-15. Practice monitor required.
All three state boards checked in a single query. The California probation with a practice monitor requirement would affect placement decisions.
Marcus would normally spend 20 to 30 minutes navigating the Florida DOH license search, the Texas TMB verification portal, and the California Medical Board's BreEZe system. One query replaced all three, and it caught a California probation that might not have surfaced if he'd only checked the states where the physician was being placed.
But Marcus also needs to fill positions. He follows up: “Find all hospitalists within 30 miles of our client hospital in El Paso. Group by facility and show which hospitals have the most. Then screen the unaffiliated ones for compliance.” One question turns into a pre-vetted candidate list with facility context, compliance status, and geographic proximity, the kind of market analysis staffing agencies normally spend days compiling from job boards and phone calls.
What's actually happening when you ask a question
Medistill isn't summarizing web pages or guessing from training data. When you ask a question, it translates your plain English into queries against structured, normalized databases that we build and maintain from primary government sources.
We pull raw data from CMS, FDA, CDC, HRSA, state boards, and dozens of other agencies. We clean it, normalize it, and keep it current. When Rachel asks to screen a provider, Medistill isn't scraping OIG's website in real time. It's querying structured data that was synced from OIG's published data files.
That matters because it's fast, reliable, and precise. It's also why Medistill can cross-reference across sources. When it checks a provider against 130+ compliance databases, it's matching on NPI numbers across all of them, not running 88 separate web searches.
There are 2,000+ datasets in total, covering hospital quality, provider enrollment, drug pricing, clinical trials, compliance and enforcement, insurance coverage, and population health. Every dataset is queryable through the same plain English interface. And every answer comes with its own custom report - tables, charts, risk scores, comparisons - generated from scratch for your specific question. There are no pre-built templates or rigid report libraries. You ask, and the output is built for exactly what you asked, including custom dashboards, charts, and reports generated on the fly.
Every follow-up crosses more data
This isn't a one-shot search box. Each follow-up builds on context and pulls in new datasets automatically. Here's what Rachel's compliance screening conversation looks like after the initial query:
“Screen Dr. Martinez, NPI 1234567893, against all enforcement databases.”
→ 130+ databases checked. Grade C, Florida board probation flagged.
“He's getting $18K in pharma payments. Who's paying him and for what?”
→ Open Payments data joined. Top payer: device company, $12K in consulting fees.
“What's he actually prescribing in Part D? Does it match who's paying him?”
→ Part D prescribing data crossed with payments. His top drug is made by his top payer.
“Are there other providers at our hospital with the same pattern, discipline plus pharma payments?”
→ Facility crosswalk identifies all 253 providers, screens each, flags 4 with similar profiles.
“Run predictive risk scores on our entire medical staff. Who looks clean today but is trending toward trouble?”
→ ML model trained on 7M providers scores every physician. 6 clean providers score above 30, high likelihood of future enforcement.
Five questions. Five different data dimensions: compliance enforcement, industry payments, drug prescribing, facility affiliations, and predictive risk modeling. Rachel didn't switch tools, re-enter the provider's information, or wait for a report to be built. Each follow-up remembered context and pulled in whatever data was needed to answer.
This is what makes Medistill different from a static reporting tool. Pre-built templates answer the questions someone anticipated when they designed the report. A conversation answers the question you actually have right now, crosses whatever datasets are relevant, and generates the dashboard, chart, or report for it on the fly.
Any format you can think of
The answer doesn't have to stay in the chat. Once you have the data you want, just say what you need:
Interactive dashboard
“Turn this into a shareable dashboard”
Live charts and tables your team can filter, sort, and explore. Share a link - anyone can view it.
PDF report
“Generate a PDF I can send to the board”
Formatted report with executive summary, charts, tables, and findings. Board-ready in seconds.
Excel spreadsheet
“Export this as an Excel file”
Raw data in columns you can pivot, filter, and feed into your own workflows.
Anything else
“Just describe what you need”
A chart for a slide deck. A summary for an email. A comparison table. A compliance checklist. If you can describe it, Medistill builds it.
The output format isn't a setting you configure. It's part of the conversation. Ask for a compliance report and you get a structured report. Ask for a chart comparing five hospitals and you get a chart. Ask for a spreadsheet you can hand to your analyst and you get a downloadable Excel file. Same data, whatever shape you need it in.
Try it with your own questions
Medistill offers a 50 free credits with full access. No credit card required to browse. Connect it to Claude and ask the healthcare data question you've been waiting on someone else to answer.
If you've ever been told “you need SQL for that” or “submit a ticket to the data team,” you don't anymore.