What problem does it solve?
Provides a centralized, authoritative reference and examples to assemble complete and correctly ordered ReactiveAgent builder chains so developers avoid missing methods, misordered calls, or incorrect parameter shapes when constructing agents.
Core Features & Use Cases
- Complete Method Catalog: Enumerates .withX() methods, expected parameter shapes, defaults, and return behaviors across identity, provider, reasoning, tools, memory, safety, cost, observability, persistence, and build APIs.
- Practical Implementations: Shows minimal and production-ready builder chains that demonstrate reliability, observability, cost routing, guardrails, and memory configuration.
- Use Case: Build a production-ready assistant that uses an external provider, a specific model, adaptive reasoning with iteration limits, enhanced memory backed by SQLite, runtime guardrails, verification, and cost tracking for safe, auditable deployments.
Quick Start
Compose an agent builder that names the agent assistant, sets provider to anthropic, selects model claude-opus-4-6, enables adaptive reasoning with maxIterations 15, configures enhanced memory at ./agent.db, turns on guardrails and verification, enables cost tracking with perRequest 0.50, and enables live observability.