Clasta — CRM Platform for Experts
The production application is not publicly available yet, but you can visit the landing page: https://getclasta.com/
Overview
Clasta is a CRM and booking platform for activity management and small businesses. The main focus is children's clubs (judo, football, dance and so on), but we kept the system general enough for any type of business, because flexibility was the first thing we argued about.
Parents use it to find and book classes for their children. The expert side is a full back office: attendance tracking, membership management, scheduling, payments and AI-assisted operations.
This was a side project I joined on a friend's suggestion. Two full-stack developers, and an architecture built around business logic, which matters more than usual in a startup where the requirements move every few weeks.
My role
Part-time full-stack developer, alongside my main job at Zenpulsar, split roughly 50/50 between backend and frontend.
- Built the entire UI component library, adapting Quasar base components to the Untitled UI design system
- Developed and maintained API endpoints for CRM features
- Implemented the business logic across the application
- Wrote unit tests (Vitest, Pytest), load tests and frontend component tests, using fixtures and mocks to keep them isolated
Key features
Attendance and memberships
- Design and track flexible membership plans
- Manage class capacity and session attendance
Integrated scheduling
- Central calendar for single events and recurring classes
- Online booking and long-term planning
Business automation
- Automated reminders, follow-ups and session confirmations
- Less admin work for the educators
AI assistant and search
- Chat-based helper for payments, invoices, payroll and marketing
- Smart search across clients and activities
Finance and operations
- Payments, invoices, marketing tools and operational metrics in one place
Technical implementation
Architecture
- Clean Architecture on the backend, layered so it stays maintainable
- Feature-Sliced Design on the frontend, organized around business features
- Business logic first, so a change in requirements lands easily in code
- npm workspaces for the monorepo
Backend
- FastAPI for REST APIs
- SQLModel for a type-safe ORM
- PostgreSQL as the main database
- Redis for caching and task queues
- MinIO for S3-compatible file storage
- Prefect for background workflows
Frontend
- Vue 3 with Quasar
- TanStack Query for data fetching and state
- Custom UI library built on the Untitled UI design system
- Chart.js for dashboards
Testing
- Vitest for frontend unit and component tests
- Pytest for backend unit tests
- Fixtures and mocks for isolated tests
- Load testing to check performance
Timeline
Tech stack
- Backend: Python, FastAPI, SQLModel, PostgreSQL, Redis, MinIO, Prefect
- Frontend: Vue 3, Quasar, TanStack Query, Chart.js
- Testing: Vitest, Pytest
- Architecture: Clean Architecture, Feature-Sliced Design