Minicor raises $3.8M seed to make legacy desktop apps callable by AI

Minicor raises $3.8M seed to make legacy desktop apps callable by AI

14 October 2026: Minicor, a startup that wants AI agents to operate old desktop programs the way they would call a modern API, has raised a $3.8 million seed round, according to a post on its own newsroom. Brad Gillespie of IA Ventures led the round, and Y Combinator and Paul Graham took part. The post is written by Faizaan Chishtie.

Not yet independently verified. The round rests on a single newsroom post by the company. We will add an independent source when one is published. We will update this when it can be confirmed, and remove this note.

The problem the company describes

Minicor says 70 percent of Fortune 500 companies still run on desktop applications that have no API. It names two ways companies cope. Robotic process automation is engineering heavy to build and to maintain. Computer-use agents are easy to set up but often slower than a person doing the same work.

The company's answer is software that turns legacy desktop programs into endpoints an AI can call like any other API. Agents learn the target software, build the automations and then keep them running. Its one-line description on the post reads: "Minicor lets AI use legacy desktop software like an API."

Who invested

Besides the lead, the post thanks a long list of funds and people. The funds named are Pioneer Fund, Robinhood Ventures, Garage Capital, Rebel Fund, Blast, SuccessVP, K5 Global, 20VC, Phosphor Capital and Metsu Capital. The angels are Austen Allred, Henri Stern, Matteo Franceschetti and Jeff Henriod.

What the website shows

Minicor's site, whose footer gives the legal name as Laminar Run, Inc., carries pages for use cases, a FAQ, security, a service level agreement and documentation, along with a GitHub link and a careers page that lists openings. The security and SLA pages suggest the product is aimed at enterprise buyers who need uptime commitments before they let software drive a production application.

Gaps in the public picture

None of the pages read states a headquarters city, a founding year or the number of employees. What the company has published is the investor list, the product description and the problem statement, so the size and customer base of the business behind the round remain unreported.

Where it sits among automation tools

The approach sits between two established categories. Robotic process automation tools record and replay clicks on a screen, and general computer-use agents drive the screen through a language model. Minicor's pitch is that its agents learn a specific program once and then expose its functions as callable endpoints, so a downstream AI system does not have to navigate the interface each time.

Company profile on TrustList: Minicor

Sources

• Minicor newsroom: Minicor raises $3.8M seed round, 14 October 2026