Roadmap Summary
This summary mirrors the current implementation plan and is intended to anchor documentation to the planned delivery order.
Phase 1: Foundation
- M1: core annotations, proxy generation, dispatch, provider contracts, exception hierarchy
- M2: structured output pipeline, schema generation, parsing, retry feedback loop
- M3: first real providers, provider routing, timeout and retry support
Phase 2: Production-Ready
- M4: observability with OpenTelemetry
- M5: standalone module and Java-friendly API surface
- M6: Spring Boot adapter and configuration model
Phase 3: Ecosystem
- M7: testing module and assertion helpers
- M8: full public documentation, live proof, publishing, and project hygiene
Phase 4: Growth
- M9: raw text streaming support (
Flow<StreamChunk>) - M10: engine-owned tool calling orchestration
- M11: Spring Boot automatic tool discovery
Phase 5: Production Hardening
Operational readiness, advanced resilience, PII masking, GraalVM support, and OTel metrics.
Future Direction
Typed orchestration and coordination for multi-service AI workflows.
Current Delivery Snapshot
- Phase 1: implemented
- Phase 2: implemented
- Phase 3: implemented
- Phase 4: implemented
- Phase 5: in progress
Documentation Implication
The docs in this repository should grow with the roadmap:
- architecture and ADRs now
- phase-grouped specs and execution tasks for committed work
- API and configuration references when modules exist
- user guides and migration notes once features are implemented
Current Documentation Coverage
- Phase 1: committed specs and tasks exist
- Phase 2: committed specs and tasks exist
- Phase 3: committed specs and tasks exist
- Phase 4: committed specs and tasks exist
- Phase 5: implementation is in progress under
SPEC-011andTASK-016 - Future direction: orchestration prototype work is underway under
SPEC-012,ADR-017, andTASK-017