Minicor Automates Desktop Software That Has No API Because Sometimes the Only Way In Is Through the Screen

The Macro: The World Still Runs on Software That Cannot Be Automated

Somewhere inside every hospital, insurance company, and logistics operation, there is a critical piece of software with no API. No webhooks. No integration layer. The only way to interact with it is through the user interface, clicking buttons and typing into fields, exactly the way a human would.

This is not ancient history. Epic, Athena, and Cerner are modern electronic health record systems used by thousands of hospitals. But their desktop clients often lack the API coverage needed for full automation. The same is true for ERP systems, automotive management software, logistics platforms, and financial services applications. These systems were designed for human operators, and they were never designed to be controlled programmatically.

Traditional RPA (robotic process automation) tried to solve this. UiPath, Automation Anywhere, and Blue Prism built bots that click through desktop interfaces. But traditional RPA is brittle. It records exact screen coordinates and pixel patterns. When the software updates its UI, adds a dialog box, or moves a button, the bot breaks. Maintaining RPA bots against evolving software is a full-time job that often costs more than the manual labor it replaced.

Minicor, backed by Y Combinator, takes a different approach. Instead of recording and replaying clicks, Minicor uses AI computer-use agents that understand the screen and adapt when things change.

The Micro: Self-Healing Bots That See and Adapt

The self-healing capability is the core differentiator. When a UI changes or an unexpected dialog appears, Minicor’s agent catches it, adapts, and keeps going. Traditional RPA breaks in this scenario. Minicor claims 93-96% click accuracy compared to 80-85% for competitors. That gap is the difference between reliable automation and constant maintenance.

Deployment flexibility covers Windows VMs, cloud, on-premise, and Citrix environments. One API call triggers a full desktop workflow. This matters because enterprise IT environments are complex. Some run on-prem. Some use Citrix for remote desktop delivery. Some are hybrid. Minicor works across all of them.

Built-in observability provides full video recordings of automation runs, Slack notifications, and detailed execution context. When something does go wrong, you can watch exactly what happened rather than debugging from error logs.

The healthcare focus is strategic. The company claims 25,000 patients per day running through their platform. Healthcare is the perfect market for this because EHR systems are mission-critical, resistant to API integration, and desperately in need of automation. SOC 2 Type II and HIPAA compliance are table stakes for healthcare software.

The founding team includes Faizaan Chishtie as CEO and Saheed Akinbile. The go-to-market timeline of “zero to production in weeks” contrasts sharply with traditional RPA deployments that take 4+ months.

The onboarding process is elegant. Users record a video of a workflow. Minicor’s platform handles the automation generation from there. No programming required. No bot configuration. Just show the platform what you need done.

The Verdict

The legacy software automation problem is enormous and persistent. These systems will not be replaced anytime soon. The only option is to work with them, and Minicor is building the best way to do that.

At 30 days: how many different desktop applications has Minicor successfully automated, and what is the success rate across first-time runs? Breadth and reliability determine market applicability.

At 60 days: what is the maintenance overhead compared to traditional RPA? If self-healing genuinely reduces the need for bot maintenance, the total cost of ownership drops dramatically.

At 90 days: are customers expanding automation from one workflow to many? The land-and-expand motion is the growth strategy, and it works when the first automation proves reliable enough to justify the second.

I think Minicor is solving the right problem with the right technology. AI computer-use agents are a generational improvement over recorded-playback RPA. The healthcare focus gives them a defensible beachhead in a market that is both large and sticky. If the self-healing claims hold up at scale, Minicor will eat traditional RPA’s lunch.