I wanted health IT market data. Definitive Healthcare quoted me $25k.

You’re a healthtech founder building your competitive landscape before a board meeting. Perhaps you need a quick read on a company before a partnership call and don’t have time to piece it together from five browser tabs. Or you’re a BD rep trying to prioritize 200 California hospitals by EHR vendor. Or you’re a sales leader mapping territory assignments and need to know which health systems are already locked into a competitor. If any of these resonate, this is for you.

I was asking the same questions. It seems like there’s a new AI scribe company being announced every week. I wanted to know which ones were actually being deployed at health systems. The market is more nuanced than a head-to-head race between Epic Ambient and a wave of startups.

Definitive Healthcare wants $25k for this data. Digging through Google, LinkedIn, press releases - all of this research took 15-20 minutes per entity. Claude and ChatGPT pulled from training data that was already months stale unless you explicitly prompt them to fetch from the web, which was still inconsistent. So I thought about where this data actually lives, and whether the paywall was real or just the default assumption. Mostly it wasn’t real. So I built my own. Varys is a free, open-source tool for mapping health IT vendors and prospecting health systems.

Ask Varys which AI scribe companies are deployed at Penn Medicine. You get Abridge, their $757M Series E, Epic integration status, cited clinical evidence, and more in under 2 minutes. The same works in reverse: give it a company name and it maps out which health systems they’ve already won. Or start from scratch and ask Varys to map a category entirely — “find Series B RCM companies selling to community hospitals” — and it returns a sourced list ready to profile.

What’s more exciting is that you can batch process a list of competitors or prospects you may already have. Point the tool at your spreadsheet and have it enriched without touching each company one by one. Unlike asking Claude or ChatGPT directly, each company gets its own isolated research pass, so there’s no context bleed, no confident-sounding guesses filling in gaps.

If you have an AI assistant, you can use it with no setup. Load the skill files from the `skills/` folder, into Claude, ChatGPT, or Gemini, and start asking questions. No code required. Use the CLI tool if you want to run batch profiles and export to CSV.

Full setup instructions are on GitHub.

In the age of AI, basic health IT market data shouldn’t require a procurement process. If you use it, find a bug, or want to see a field added — open an issue or reach out.