Show HN: I made Python library for composition-first AI programming (github.com)

Funcai, a newly announced library for composing LLM (Large Language Model) interactions, presents a functional programming alternative to frameworks like LangChain. This approach allows developers to structure AI dialogues as functions rather than relying on complex inheritance chains. The core of Funcai's architecture is the Dialogue construct, which enables interactions with various tools and interpreters. Developers can easily create interactions, such as evaluating mathematical expressions or managing item storage, using clear, composable code. The library operates lazily, executing only when explicitly awaited, and employs advanced error handling techniques that simplify managing failure cases. The significance of Funcai for the AI/ML community lies in its innovative design principles that prioritize modularity, ease of use, and extensibility. By abstracting dialogue as a functional operation and introducing combinators to manage complex workflows, Funcai encourages cleaner code and fosters experimentation with AI interactions. Built on Python 3.14+ with native generics, it offers developers the flexibility to create custom provider and agent classes, making it a powerful tool for building sophisticated AI applications with minimal overhead.