Huntr
5,000+ users in three months
Built a Bubble-powered AI job platform with resume tailoring, application tracking and a browser extension — from broken prototype to 5,000+ users.
- 0+
- users
- 0%
- faster apply
What was blocking progress
- Broken prototype from previous developer
- AI integration not production-ready
- No proper data architecture
The situation
Huntr's founder had a clear product vision — an AI-powered job search platform that tailors resumes to each listing — but the previous developer left them with a broken prototype that couldn't handle real users.
The prototype had AI calls hardcoded with no error handling, job data stored in flat text fields with no relational structure, and no way to track where a user was in their application pipeline. The founder needed a Bubble developer who could rebuild it as a production SaaS.
Bubble was the right platform because the product needed rapid iteration on user flows — the job search process has many steps and each needed testing with real users. Bubble's visual editor and workflow system let us ship new features daily during the build phase.
What we built
We rebuilt the Bubble app from the broken prototype, adding proper data architecture, reliable AI integration and a clean user experience.
Application tracking data architecture
We designed custom data types for Jobs, Applications, Stages and Resumes with relational links between them. Each application moves through pipeline stages (saved, applied, interviewing, offered, rejected) with timestamp tracking. "Do a search for" queries with stage constraints let users filter their pipeline view without loading every application.
AI resume tailoring integration
The OpenAI integration was rebuilt using the API Connector plugin with proper error handling, rate limiting and response validation. When a user selects a job listing, a Bubble workflow extracts the listing's keywords, sends them with the user's base resume to the OpenAI API, and returns a tailored version. Scheduled API workflows handle the async processing so the UI stays responsive.
Job tracking dashboard
The main dashboard is a Kanban-style pipeline view built with Bubble's repeating groups and custom states. Users drag applications between stages, and backend workflows update the status and trigger follow-up reminders through scheduled API workflows. Conditional formatting highlights applications that need attention.
Performance at scale
With 5,000+ users, the Bubble database needed careful optimisation. We implemented pagination on all list views, used "Do a search for" with tight constraints on every query, and deferred heavy elements with "Only when" conditions. The dashboard loads in under 2 seconds even for users with hundreds of tracked applications.
- Relational data types for job applications with pipeline stage tracking
- OpenAI resume tailoring via the API Connector plugin
- Kanban-style dashboard with drag-and-drop and conditional formatting
- Performance optimisation for 5,000+ users
Technical challenge
The AI workflow had to take a job listing, extract keywords, compare them against the user's resume, and generate tailored versions — all through the OpenAI API with proper rate limiting and error handling.
The result
Users cut application time by 60%. 2x more interview callbacks reported. The Bubble app scaled to 5,000+ users in the first three months with clean page loads.
The AI integration handles thousands of resume tailoring requests reliably through the API Connector. The founder's team continues to iterate on features using the Bubble editor and documented workflow architecture.
Have a Bubble AI prototype that needs production engineering?
Next case study
Workello Hiring assessment platform