MindFort Runs 100+ AI Pen Testing Agents Without Their Previous $10k AWS Bill

For a security platform that deploys specialized AI agents to perform continuous penetration testing, infrastructure costs can make or break the business model. MindFort allows companies to run daily pen tests with the same fidelity as human testers, replacing quarterly or annual assessments that leave vulnerabilities exposed for months.

Six months ago, they faced a seemingly impossible equation: maintain 100+ agent workers available 24/7 to meet their 4-hour target, while staying profitable serving both startups and enterprise customers.

"We basically need to support concurrency so that our wait times are good. To maintain our target of getting pen tests done within four hours—which is enormously fast compared to a human—we have to have 100 workers available per customer."

The challenge was existential. Each worker requires 4 gigabytes of RAM and 4 CPUs, and they must be constantly running, registered with Temporal, waiting to pick up jobs. Traditional cloud providers would charge for this idle capacity continuously.

"In Amazon, our bill would be prohibitively expensive to leave workers running without auto-scaling on. You're talking about reserving 4 gigabytes and 4 CPUs per worker indefinitely. The fleet has to be always on and always running."

Running 100+ workers on AWS would cost approximately $10,000 per month—more than most customers were paying for the service itself. The math simply didn't work.

MindFort discovered Railway's unique pricing model perfectly aligned with their architecture needs. They could maintain massive capacity without paying for idle compute.

"We run all of our agent workers in Railway. All of them. Since we're only charged for compute that's used, we can have that capacity sitting there, but we're not paying for it."

The platform transformed not just their production infrastructure, but their entire development workflow. MindFort's team doesn't develop locally—they develop directly on Railway.

"This is ridiculous and it's going to sound foolish, but we don't develop locally. We develop on Railway. It would be a whole pain in the ass to deploy our services directly via Kubernetes on everyone's laptops."

Railway's deployment speed made this unconventional approach viable. Engineers ship changes directly to their shared dev environment, testing in real-time without local development overhead.

"Our development environment is also in Railway and we do core application development by shipping up a change. Thank God you guys deploy fast. There's no local development that happens."

The visual canvas and branch switching capabilities streamlined their workflow. The team works in a shared dev environment, quickly switching branches to test features before merging.

"One of the benefits of Railway is that we can quickly switch branches, deploy the service we're testing in dev, and then we merge up to our dev branch, do one final QA and then dev gets merged into main."

MindFort runs sophisticated infrastructure on Railway: multiple groups of AI agents, headless browsers, various security tools, Temporal for orchestration, and an AI gateway server normalizing model providers to OpenAI API calls. Each worker runs many agents in parallel, executing Python code, using tools, and processing results.

• Infrastructure costs dropped 97%, from an estimated $10,000 monthly on AWS to just $200-300 on Railway. This dramatic reduction happened while actually improving performance and maintaining their aggressive 4-hour target.
• 100+ workers run continuously without bankrupting the company, supporting all customers with 100s of workers each to maintain concurrency. The always-on fleet that would have been prohibitively expensive elsewhere became sustainable on Railway.

"If we didn't have Railway, we would literally be spending inordinate amounts of money."

Enterprise customers and startups alike receive pen tests in 4 hours instead of the industry-standard week. The platform has grown rapidly in six months, serving everyone from startups to enterprises in finance, technology, and healthcare requiring private deployments.

"We would not have been able to develop as quickly and deploy such scale without Railway. Our business would be behind where we are now if we didn't have the ability to use Railway, develop quickly, iterate quickly, and launch with the capacity that would otherwise be extremely prohibitively expensive and complex to manage somewhere else, full stop."

Looking forward, MindFort continues to migrate more infrastructure to Railway, moving their entire self-service cloud platform from Amazon while maintaining private deployments for enterprise customers requiring on-premise solutions.