Inspiration
Security feels backwards: teams ship fast, then slow down later to hunt bugs. We also noticed something else—one AI working alone is still bottlenecked, just like one person. So we asked: what if we used AI the way real teams work—multiple “brains” at once—so security moves at the speed of building?
What it does
Interlinked is a fast security assistant that uses multiple AIs at the same time to scan a codebase, spot weak points, and suggest fixes. Instead of waiting hours to find what’s wrong, you get the “what” and the “how to fix it” quickly—so teams can keep shipping.
How we built it
We designed Interlinked like a small security team:
One AI keeps the overall plan and prioritizes what matters most
Other AIs check different parts of the code in parallel
Their results get combined into one clear summary with suggested fixes
The core idea is simple: if you split the work, you cut the waiting.
Too many warnings: early versions flagged everything. We had to focus on what’s actually risky.
Keeping results consistent: different AIs can disagree, so we worked on merging answers into one clean recommendation.
Speed vs trust: going fast is easy; being fast and reliable is harder.
Making it demo-friendly: we needed a smooth flow a judge can understand in under a minute.
Accomplishments that we're proud of
A multi-AI workflow that feels like a “security team,” not a chatbot
Clear, readable results that don’t require security expertise
A fast demo flow that shows real value immediately
A product story that’s simple: faster security, fewer surprises
What we learned
Coordination matters as much as intelligence—multiple AIs only help if they stay aligned.
The best demos aren’t the most complex—they’re the most understandable.
“Findings” aren’t enough. People want fixes and clear next steps.
What's next for Interlinked
Add a verification step so suggested fixes can be checked automatically
Expand coverage to more common stacks and frameworks
Improve prioritization so the first results are always the most important
Turn it into a simple workflow: scan → top risks → suggested fixes → verify
Log in or sign up for Devpost to join the conversation.