disclaimer: no deployment URL because we can't afford credits :(
Inspiration
IRONHYDE is a tool that automatically performs black-box testing on your web apps using AI agents powered by the Model Context Protocol. 🤖
Only 17% of open-source projects have a testing suite despite it costing 6x more to fix a bug found during production that one during development. Testing, however, is mundane at best and also very time-consuming.
Prioritizing ease of use & intuitiveness, we wanted to create a product that would allow students, vibecoders, and small teams that lack the time and resources to rigorously test their projects, giving them the confidence to build fast without breaking things 🚀
What it does
Give it a URL, and IRONHYDE will:
- Identify all routes for a given domain
- Inspect the DOM and identify an exhaustive list of possible user interactions
- Execute each interaction to observe whether or not any erroneous behaviour arises
- Lets you know what went wrong, and how to replicate it!
How we built it
Built on top of a foundation provided by langchain & playwright's MCP server, we created uAgents (thanks Fetch.ai ❤️) that performed various roles such as scouring for page routes, mapping our interactions, and executing them.
Treating these agents as though they were microservices, our frontend orchestrates the whole testing flow and streams a live output of each interaction for the user.
Challenges we ran into
We often hit usage limits due to the large amount context we were working with, and we learnt to chunk our messages as a workaround. State management in react was extremely tough to juggle due to the dynamic nature of the data we were working with (nested interactions etc.), but we figured it out!
Accomplishments that we're proud of
We believe this is a product that would be extremly useful to the public, and we're glad we saw that vision through. Building a frontend that was sleek, intuitive and user-friendly was a challenge, but we're proud of how our app looks :)
What we learned
The future is here! With the rise of vibe-coded apps and the pressure to ship fast, writing reliable, scalable apps are ever more important
What's next for IRONHYDE
✅ Integration with Github Actions ✅ Infrastructure to handle multiple requests by spinning up multiple VMs ✅ Contributing to playwright-mcp to add on additional interactions ✅ Hybrid interface for users to execute human-only tasks (like CAPTCHAs) ✅ Ability to pass parameters to an interaction (authentication etc.)
Log in or sign up for Devpost to join the conversation.