Minicor

At Minicor, we’re tackling the gross but fun problem of getting AI into legacy systems. Our customers are AI companies who sell to enterprise customers (think hospitals, warehouses, HVAC companies, car dealerships, hotels, …) who all are dying to use AI but aren’t able to because their critical systems are all super old Windows desktop applications with no APIs that run on-premises.

The only way to “reliably” connect to those systems is to build UI automations (RPAs) that execute sequences of clicks and keystrokes to complete a task like a human would. These are inherently extremely brittle because a single pixel change can break critical tasks like inputting medical record details. On average, before using Minicor, our customers report error rates north of 30% for their RPA integrations - where each failure requires one of their engineers to physically log into the desktop running the automation and debug/diagnose the issue BY HAND. This becomes a critical issue at scale where 1000+ automation runs per day mean hundreds of engineering support tickets.

Minicor lets developers build Windows desktop automations at scale. Engineers use our MCP + claude code/cursor to build Desktop RPAs and store them as step based workflows that can be triggered via API. Each workflow execution on Minicor stores video replays and logs, so when one fails a Minicor agent can diagnose and take actions to fix the issue or notify a human.

We are a small team of multi time founders and engineers who love working with customers, distributed systems, and most importantly really really old disgusting horrible legacy software.

We have a lot of fun as a team, we respect and trust each other and uphold a strict "no ego" policy at the company.

Currently growing super fast, things are constantly breaking, and are looking for people to help urgently!

If you find it FUN to connect disparate systems together at scale, you'll love working at Minicor. We interact with critical software older than us regularly, and a big part of our job is being crafty about building solutions to read and write to them at scale with low latency.

Core components of our tech: our workflow orchestrator, our MCP, and our windows desktop service. On top of these are many other systems which all work together to ensure we survive multiple gigabytes of data being transferred on our platform per week, and thousands of RPA runs per day.

We (hilariously) have deployments on GCP, Azure, and AWS.

Some other frameworks/tools we use (outside of Claude Code + Cursor etc...): PSQL, Redis, Azure VMs, Nest.js, Next.js, Spring Boot, Flask.