Inspiration

Structural steel and fabrication work involves moving between drawings, CAD software, supplier catalogues, and quote requests. We wanted to make those steps easier to access by bringing them into one AI-assisted workspace built for engineers and fabricators.

What it does

Metra lets users describe a part in plain English and generate a downloadable STEP model with an in-browser 3D preview. They can also upload CAD files or images for AI-assisted drawing analysis, browse and get recommendations from a steel components catalog, request quotes, and ask the Metra AI assistant questions.

How we built it

We built the interface with Next.js and React, then used Convex for the database, file storage, backend actions, and static hosting. Zoo Dev / KittyCAD powers text-to-CAD generation; Azure OpenAI handles chat, drawing analysis, and product recommendations. Three.js and OpenCascade.js provide browser-based 3D previews, and Upstash Redis supports rate limiting.

Challenges we ran into

A key challenge was connecting several demanding workflows—CAD generation, file uploads and analysis, and product discovery—while keeping API credentials on the server. We also had to make generated or uploaded geometry useful in the browser, which meant supporting file parsing and 3D previews alongside a statically hosted frontend.

Accomplishments that we're proud of

We brought part generation, drawing analysis, product discovery, and quote requests together in one working platform. We’re especially proud that a plain-language part description can lead to a downloadable STEP file and an interactive 3D preview.

What we learned

We learned how to use Convex as more than a database: it can support the backend actions, file storage, and hosting behind a single app. We also learned that AI features are most useful when they connect to practical engineering tasks, such as inspecting a drawing or finding a relevant component.

What's next for Metra

We’d like to improve the accuracy and usefulness of generated CAD and drawing analysis, expand the product catalogue and recommendations, and make it easier for teams to move from a design to a supplier quote.

Built With

Share this project:

Updates

Submission history