Features
Everything you need to build production-ready applications from visual models
Fractal Studio
A visual modeling environment where you define your entire domain. No code required at this stage, just structured thinking about your business logic.
- Entities with fields, relations, and value objects
- Commands that mutate state and emit events
- Queries for reading data with filtering and pagination
- Services & Repositories with swappable adapter implementations
- RBAC permissions defined per role, per operation
What Gets Generated
Not scaffolding. Not boilerplate. Production-ready code that follows best practices.
FastAPI Backend
Full CQRS architecture with command handlers, query handlers, event emitters, and API routes. Type-safe with Pydantic models.
SvelteKit Frontend
Working UI out of the box: forms, lists, detail views. Custom frontend components can be specified per field.
Test Suite
API tests, service tests, all using dummy adapters. Advanced RBAC testing asserts every role against every endpoint.
Deployment Config
Dockerfiles, GitHub Actions workflows, GCP Cloud Build and Cloud Run configs. Ready for CI/CD.
GitHub Integration
Generated code commits directly to a new branch via the Fractal GitHub App. Merge when ready.
Fractal Framework
Pip-installable runtime libraries: specifications, repositories, commands, roles, tokens, and projections.
Branch-Based Regeneration
The key to keeping generated code maintainable: every regeneration creates a new Git branch. Your custom code on main is never touched.
- Custom code preserved in the
custom/folder - Override generated code without losing customizations
- Review before merging so you're always in control
Role-Based Access Control
Define permissions per role, per operation, right in the model. The generated code enforces these rules at the API layer, and generates tests that verify every role/endpoint combination.
- Record-level permissions
- Auto-generated RBAC tests
- Introspection API
Dependency Injection
Adapters are declared in Studio, then selected at runtime via environment variables. The same codebase runs against different backends without code changes.
Databases
MongoDB, PostgreSQL, Firestore, DuckDB
ORMs
SQLAlchemy, Django, Elasticsearch
Testing
Local file storage, in-memory adapters
Define service functions with typed parameters, then implement with your business logic later
Fractal Framework
Open source Python libraries that power the generated code. Use them standalone or as part of a Fractal Forge project.