HomeToolsMCPHow It WorksStoriesPhilosophyArchitectureStar on GitHub
All Tools
B
Dev ToolsFreemiumOpen Source

BAML

Type-safe language and tooling for reliable AI workflows

Apache-2.0

ABOUT

Prompt-heavy AI applications often rely on fragile string templates and ad hoc JSON parsing. BAML gives teams a typed, testable interface for prompts and model calls so agent workflows can produce structured outputs reliably across languages, models, and deployment environments.

INSTALL
pip install baml-py baml-cli init

INTEGRATION GUIDE

1. Build type-safe AI agents and LLM workflows with generated clients 2. Extract structured data from documents using validated model outputs 3. Test prompts in development and CI before shipping production changes 4. Add retries, fallbacks, streaming, and model switching to AI features

TAGS

prompt-engineeringstructured-outputstype-safetyai-agentstestingllm-framework