Functional Api With Task And Entrypoint Decorators

via “functional task-based workflow definition with @task and @entrypoint decorators”

Graph-based framework for stateful multi-agent LLM applications with cycles and persistence.

Unique: Decorator-based functional API that automatically constructs StateGraph under the hood, enabling implicit state threading and dependency injection while maintaining full Pregel execution semantics

vs others: More concise than explicit StateGraph for simple workflows, but less transparent than imperative code for complex control flow

via “functional decorator-based task definition with @task and @entrypoint”

Build resilient language agents as graphs.

Unique: Uses Python function introspection and type hints to automatically infer state channel bindings and merge semantics, eliminating manual edge/channel declarations. The @entrypoint decorator compiles decorated functions into a fully executable graph without explicit StateGraph construction.

vs others: Offers a more Pythonic, decorator-driven alternative to explicit graph construction while maintaining full compatibility with Pregel execution, reducing boilerplate for simple workflows compared to StateGraph while preserving power for complex cases.

via “functional api with @task and @entrypoint decorators”

Building stateful, multi-actor applications with LLMs

Unique: Implements a functional programming interface with @task and @entrypoint decorators that automatically infer state schema from function signatures and construct implicit graphs, reducing boilerplate for simple workflows while maintaining access to full StateGraph capabilities.

vs others: More concise than explicit StateGraph definitions for simple workflows while remaining more explicit than implicit agent frameworks, enabling developers to choose between functional and declarative styles.