BAML

BAML: The First Language for Building AI Agents

What is BAML? BAML, or Basically A Made-Up Language, is an open-source toolkit designed to empower developers in building type-safe and reliable AI applications. Similar to how TypeScript enhances JavaScript, BAML aims to improve the reliability of AI pipelines by providing a structured approach to AI development.

How does BAML work?

BAML helps developers define, test, and deploy AI agents with confidence. It works with every LLM provider and supports multiple languages, including Python, TypeScript, Ruby, and Go. BAML transforms AI development through a simple four-step process:

1. Define Prompts Functions: Create prompts using BAML within your IDE. BAML is compatible with tools like Cursor and Claude and offers a VSCode extension.
2. Test Prompts Functions: Test your prompt functions in VSCode or your preferred editor. You can also use the baml-cli test in CI/CD pipelines.
3. Call Prompts Functions: Generate native functions from BAML functions in languages like Python, TypeScript, Ruby, and Go.
4. Deploy Your Agent: Use BAML in any way you want, as it generates native code that integrates seamlessly into your existing infrastructure.

Key Features and Benefits

• Type Safety: Define AI interfaces confidently with BAML schemas that automatically generate TypeScript types. It ensures structured outputs in formats like JSON, XML, and YAML.
• Complete Development Workflow: From defining prompts to testing and deployment, BAML offers a comprehensive workflow to streamline AI application development.
• Automatic Retry and Fallback: Automatically retry failed requests and provide fallback responses when errors occur, ensuring more reliable AI interactions.
• Multi-Cloud Deployment: Supports deployment across multiple cloud platforms like AWS Lambda, Vercel, Google Cloud, Azure Functions, and Railway.
• Testing in CI/CD: Test agents in Continuous Integration and Continuous Deployment pipelines to ensure they perform as expected.

Trusted by Developers

BAML is used and trusted by developers at leading companies. Users praise its ability to provide structured data from LLMs, improve iteration speed and quality, and offer strong type guarantees. Here's what some users are saying:

• Adam Gitzes from Amazon: "BAML is amazing. I've used it in Python and Typescript. It's a game changer."
• Jason Fan from Finic.ai: "Just set up baml for my project, 10/10 experience and much faster than langchain."
• Ray del Vecchio from Cerebral Valley: "It's amazing!! Was able to cut down my tokens and time-to-first-token significantly without compromising results."

Why Choose BAML?

• Enhanced Reliability: BAML helps build AI applications with improved reliability by providing type safety and structured interfaces.
• Simplified Development: It streamlines the AI development process with tools for defining, testing, and deploying AI agents.
• Flexibility: BAML is compatible with multiple LLM providers and supports multiple programming languages, offering developers the flexibility they need.

Who is BAML for?

BAML is for developers who want to build AI applications with confidence and reliability. It is particularly useful for those who:

• Need structured outputs from LLMs.
• Want to improve the iteration speed and quality of their AI projects.
• Desire strong type guarantees for their AI interfaces.

How to Get Started with BAML

To start building type-safe AI applications with BAML, follow these steps:

1. Install BAML using uv add baml-py && uv run baml-cli init.
2. Define your prompts and functions using BAML schemas.
3. Test your agents using baml-cli test.
4. Deploy your agent to your preferred platform.

With BAML, developers can build AI applications with confidence, knowing their AI pipelines are reliable and type-safe.