Firefly Framework
Enterprise service patterns, integration, data access, security, and operational visibility.
The open-source frameworks we build our services with.
The Firefly frameworks support the applications, integrations, AI tasks, and processes behind our services. Each project has its own scope and releases.
Choose the right foundation for the workload
jafly, pyfly, larafly, rsfly, and netfly
Application foundations selected according to the workload, team, and integration requirements.
Firefly Agentic
Tools for structured AI tasks, permitted tool use, retrieval, and approval points.
Firefly Weave
Business processes that cross systems and people, defined once and followed case by case.
Framework choice follows service requirements
We assess release maturity, resilience, security, integration, and operational support before selecting a component. Capabilities are not assumed to be identical across language implementations.
Explore the projects directly
The project documentation and repositories describe supported features, releases, and licenses. Visit fireflyframework.org and the Firefly open-source organization on GitHub.
jafly
Build services with explicit domain boundaries. The Java modules provide command/query handling, reactive data access, API contracts, events, orchestration, security and observability.
Spring Boot / WebFlux · PostgreSQL / R2DBC · CQRS · Saga / TCC
Technical documentationpyfly
A common structure for Python services: dependency injection, domain models, APIs, data access, and integration alongside your existing technology.
Async Python · Domain services · Data adapters · Security
Technical documentationFirefly Agentic
Build agents with typed outputs, tools, retrieval, and approval points; the application controls which actions they may take.
Pydantic AI · Tool approvals · Knowledge retrieval · Pipelines
Technical documentationFirefly Weave
Versioned workflows with stored progress, human tasks, and independent workers, run through APIs, a CLI, an SDK, and a visual Studio.
PostgreSQL state · Waits and signals · Human review · AI workers
Technical documentationThe ecosystem also includes native PHP, Rust, and .NET projects, and a Data Science project for model training and evaluation. Each has its own implementation and release lifecycle. We select the runtime, persistence and adapters for the workload; capabilities are not identical across languages.
Your first operation
Bring us one operation.
Tell us the process, the volume, and what “done” looks like. We will come back with a scope, a deployment plan, and a proposal.