Integrate BAML to enable structured outputs from LLMs

[bucurdavid]

Describe the solution you'd like

I would like to integrate BAML (a domain-specific language for generating structured outputs from LLMs) into eliza workflow. BAML provides an excellent developer experience with features like type-safe outputs, flexibility across different LLMs and languages, and a robust online playground for testing prompts. It allows viewing and running prompts directly within the editor, similar to Markdown Preview, without additional setup.

Additional context

BAML enables building reliable agents and chatbots with Retrieval-Augmented Generation (RAG) capabilities, extracting data from PDFs, and more. It offers a state-of-the-art structured output experience that even outperforms OpenAI with their models and supports open-source models. The BAML VSCode playground provides an amazingly fast developer experience for prompting, with fully type-safe outputs even when streaming structured data. For more information, please refer to the BAML products, guide, playground, examples, reference, and motivation sections.

https://github.com/BoundaryML/baml