Abel
PrototypeA typed AI agent framework that makes multi-step LLM pipelines debuggable, composable, and auditable.
Built a production-grade AI agent framework from scratch
- ›Designed a typed execution graph for multi-step LLM workflows
- ›Supports OpenAI, Anthropic, and local models via a unified adapter pattern
- ›Makes AI pipelines debuggable, auditable, and testable — not just functional