base on The AI framework that adds the engineering to prompt engineering (Python/TS/Ruby/Java/C#/Rust/Go compatible) <div align="center"> <a href="https://boundaryml.com?utm_source=github" target="_blank" rel="noopener noreferrer"> <picture> <source media="(prefers-color-scheme: dark)" srcset="fern/assets/baml-lamb-white.png"> <img src="fern/assets/baml-lamb-white.png" height="64" id="top"> </picture> </a> </div> <div align="center"> [](https://pypi.org/project/baml-py/) ## BAML: Basically a Made-up Language <h4> [Homepage](https://www.boundaryml.com/) | [Docs](https://docs.boundaryml.com) | [BAML AI Chat](https://www.boundaryml.com/chat) | [Discord](https://discord.gg/BTNBeXGuaS) </h4> </div> BAML is a simple prompting language for building reliable **AI workflows and agents**. BAML helps you turn prompt engineering into _schema engineering_ -- where you mostly focus on the models of your prompt -- to get more reliable outputs. You don't need to write your whole app in BAML, only the prompts! You can wire-up your LLM Functions in any language of your choice! See our quickstarts for [Python](https://docs.boundaryml.com/guide/installation-language/python), [TypeScript](https://docs.boundaryml.com/guide/installation-language/typescript), [Ruby](https://docs.boundaryml.com/guide/installation-language/ruby) and [Go, and more](https://docs.boundaryml.com/guide/installation-language/rest-api-other-languages). BAML comes with all batteries included -- with full typesafety, streaming, retries, wide model support, even when they don't support native [tool-calling APIs](#enable-reliable-tool-calling-with-any-model-even-when-they-dont-support-it) **Try BAML**: [Prompt Fiddle](https://www.promptfiddle.com) • [Interactive App Examples](https://baml-examples.vercel.app/) ## The core BAML principle: LLM Prompts are functions The fundamental building block in BAML is a function. Every prompt is a function that takes in parameters and returns a type. ```rust function ChatAgent(message: Message[], tone: "happy" | "sad") -> string ``` Every function additionally defines which models it uses and what its prompt is. ```rust function ChatAgent(message: Message[], tone: "happy" | "sad") -> StopTool | ReplyTool { client "openai/gpt-4o-mini" prompt #" Be a {{ tone }} bot. {{ ctx.output_format }} {% for m in message %} {{ _.role(m.role) }} {{ m.content }} {% endfor %} "# } class Message { role string content string } class ReplyTool { response string } class StopTool { action "stop" @description(#" when it might be a good time to end the conversation "#) } ``` ## BAML Functions can be called from any language Below we call the ChatAgent function we defined in BAML through Python. BAML's Rust compiler generates a "baml_client" to access and call them. ```python from baml_client import b from baml_client.types import Message, StopTool messages = [Message(role="assistant", content="How can I help?")] while True: print(messages[-1].content) user_reply = input() messages.append(Message(role="user", content=user_reply)) tool = b.ChatAgent(messages, "happy") if isinstance(tool, StopTool): print("Goodbye!") break else: messages.append(Message(role="assistant", content=tool.response)) ``` You can write any kind of agent or workflow using chained BAML functions. An agent is a while loop that calls a Chat BAML Function with some state. And if you need to stream, add a couple more lines: ```python stream = b.stream.ChatAgent(messages, "happy") # partial is a Partial type with all Optional fields for tool in stream: if isinstance(tool, StopTool): ... final = stream.get_final_response() ``` And get fully type-safe outputs for each chunk in the stream. ## Test prompts 10x faster, right in your IDE BAML comes with native tooling for VSCode (jetbrains + neovim coming soon). **Visualize full prompt (including any multi-modal assets), and the API request**. BAML gives you full transparency and control of the prompt.  **Using AI is all about iteration speed.** If testing your pipeline takes 2 minutes, in 20 minutes, you can only test 10 ideas. If testing your pipeline took 5 seconds, in 20 minutes, you can test 240 ideas.  No need to login to any website to test things, and no need to copy JSON blobs around. ## Enable reliable tool-calling with any model BAML works even when the models don't support native tool-calling APIs. We created the SAP (schema-aligned parsing) algorithm to support the flexible outputs LLMs can provide, like markdown within a JSON blob or chain-of-thought prior to answering. [Read more about SAP](https://www.boundaryml.com/blog/schema-aligned-parsing) With BAML, your structured outputs work in Day-1 of a model release. No need to figure out whether a model supports parallel tool calls, or whether it supports recursive schemas, or `anyOf` or `oneOf` etc. See it in action with: **[Deepseek-R1](https://www.boundaryml.com/blog/deepseek-r1-function-calling)** and [OpenAI O1](https://www.boundaryml.com/blog/openai-o1). ## Switch from 100s of models in a couple lines ```diff function Extract() -> Resume { + client openai/o3-mini prompt #" .... "# } ``` [Retry policies](https://docs.boundaryml.com/ref/llm-client-strategies/retry-policy) • [fallbacks](https://docs.boundaryml.com/ref/llm-client-strategies/fallback) • [model rotations](https://docs.boundaryml.com/ref/llm-client-strategies/round-robin). All statically defined.  Want to do pick models at runtime? Check out the [Client Registry](https://docs.boundaryml.com/guide/baml-advanced/llm-client-registry). We support: [OpenAI](https://docs.boundaryml.com/ref/llm-client-providers/open-ai) • [Anthropic](https://docs.boundaryml.com/ref/llm-client-providers/anthropic) • [Gemini](https://docs.boundaryml.com/ref/llm-client-providers/google-ai-gemini) • [Vertex](https://docs.boundaryml.com/ref/llm-client-providers/google-vertex) • [Bedrock](https://docs.boundaryml.com/ref/llm-client-providers/aws-bedrock) • [Azure OpenAI](https://docs.boundaryml.com/ref/llm-client-providers/open-ai-from-azure) • [Anything OpenAI Compatible](https://docs.boundaryml.com/ref/llm-client-providers/openai-generic) ([Ollama](https://docs.boundaryml.com/ref/llm-client-providers/openai-generic-ollama), [OpenRouter](https://docs.boundaryml.com/ref/llm-client-providers/openai-generic-open-router), [VLLM](https://docs.boundaryml.com/ref/llm-client-providers/openai-generic-v-llm), [LMStudio](https://docs.boundaryml.com/ref/llm-client-providers/openai-generic-lm-studio), [TogetherAI](https://docs.boundaryml.com/ref/llm-client-providers/openai-generic-together-ai), and more) ## Build beautiful streaming UIs BAML generates a ton of utilities for NextJS, Python (and any language) to make streaming UIs easy.  BAML's streaming interfaces are fully type-safe. Check out the [Streaming Docs](https://docs.boundaryml.com/guide/baml-basics/streaming), and our [React hooks](https://docs.boundaryml.com/guide/framework-integration/react-next-js/quick-start) ## Fully Open-Source, and offline - 100% open-source (Apache 2) - 100% private. AGI will not require an internet connection, neither will BAML - No network requests beyond model calls you explicitly set - Not stored or used for any training data - BAML files can be saved locally on your machine and checked into Github for easy diffs. - Built in Rust. So fast, you can't even tell it's there. ## BAML's Design Philosophy Everything is fair game when making new syntax. If you can code it, it can be yours. This is our design philosophy to help restrict ideas: - **1:** Avoid invention when possible - Yes, prompts need versioning — we have a great versioning tool: git - Yes, you need to save prompts — we have a great storage tool: filesystems - **2:** Any file editor and any terminal should be enough to use it - **3:** Be fast - **4:** A first year university student should be able to understand it ## Why a new programming language We used to write websites like this: ```python def home(): return "<button onclick=\"() => alert(\\\"hello!\\\")\">Click</button>" ``` And now we do this: ```jsx function Home() { return <button onClick={() => setCount(prev => prev + 1)}> {count} clicks! </button> } ``` New syntax can be incredible at expressing new ideas. Plus the idea of maintaining hundreds of f-strings for prompts kind of disgusts us 🤮. Strings are bad for maintainable codebases. We prefer structured strings. The goal of BAML is to give you the expressiveness of English, but the structure of code. Full [blog post](https://www.boundaryml.com/blog/ai-agents-need-new-syntax) by us. ## Conclusion As models get better, we'll continue expecting even more out of them. But what will never change is that we'll want a way to write maintainable code that uses those models. The current way we all just assemble strings is very reminiscent of the early days PHP/HTML soup in web development. We hope some of the ideas we shared today can make a tiny dent in helping us all shape the way we all code tomorrow. ## FAQ | | | | - | - | | Do I need to write my whole app in BAML? | Nope, only the prompts! BAML translates definitions into the language of your choice! [Python](https://docs.boundaryml.com/guide/installation-language/python), [TypeScript](https://docs.boundaryml.com/guide/installation-language/typescript), [Ruby](https://docs.boundaryml.com/guide/installation-language/ruby) and [more](https://docs.boundaryml.com/guide/installation-language/rest-api-other-languages). | | Is BAML stable? | Yes, many companies use it in production! We ship updates weekly! | | Why a new language? | [Jump to section](#why-a-new-programming-language) | ## Contributing Checkout our [guide on getting started](/CONTRIBUTING.md) --- Made with ❤️ by Boundary HQ in Seattle, WA P.S. We're hiring for software engineers that love rust. [Email us]([email protected]) or reach out on [discord](https://discord.gg/ENtBB6kkXH)! <div align="left" style="align-items: left;"> <a href="#top"> <img src="https://img.shields.io/badge/Back%20to%20Top-000000?style=for-the-badge&logo=github&logoColor=white" alt="Back to Top"> </a> </div> <img src="https://imgs.xkcd.com/comics/standards.png" alt_text="hi" /> ", Assign "at most 3 tags" to the expected json: {"id":"10887","tags":[]} "only from the tags list I provide: [{"id":77,"name":"3d"},{"id":89,"name":"agent"},{"id":17,"name":"ai"},{"id":54,"name":"algorithm"},{"id":24,"name":"api"},{"id":44,"name":"authentication"},{"id":3,"name":"aws"},{"id":27,"name":"backend"},{"id":60,"name":"benchmark"},{"id":72,"name":"best-practices"},{"id":39,"name":"bitcoin"},{"id":37,"name":"blockchain"},{"id":1,"name":"blog"},{"id":45,"name":"bundler"},{"id":58,"name":"cache"},{"id":21,"name":"chat"},{"id":49,"name":"cicd"},{"id":4,"name":"cli"},{"id":64,"name":"cloud-native"},{"id":48,"name":"cms"},{"id":61,"name":"compiler"},{"id":68,"name":"containerization"},{"id":92,"name":"crm"},{"id":34,"name":"data"},{"id":47,"name":"database"},{"id":8,"name":"declarative-gui "},{"id":9,"name":"deploy-tool"},{"id":53,"name":"desktop-app"},{"id":6,"name":"dev-exp-lib"},{"id":59,"name":"dev-tool"},{"id":13,"name":"ecommerce"},{"id":26,"name":"editor"},{"id":66,"name":"emulator"},{"id":62,"name":"filesystem"},{"id":80,"name":"finance"},{"id":15,"name":"firmware"},{"id":73,"name":"for-fun"},{"id":2,"name":"framework"},{"id":11,"name":"frontend"},{"id":22,"name":"game"},{"id":81,"name":"game-engine "},{"id":23,"name":"graphql"},{"id":84,"name":"gui"},{"id":91,"name":"http"},{"id":5,"name":"http-client"},{"id":51,"name":"iac"},{"id":30,"name":"ide"},{"id":78,"name":"iot"},{"id":40,"name":"json"},{"id":83,"name":"julian"},{"id":38,"name":"k8s"},{"id":31,"name":"language"},{"id":10,"name":"learning-resource"},{"id":33,"name":"lib"},{"id":41,"name":"linter"},{"id":28,"name":"lms"},{"id":16,"name":"logging"},{"id":76,"name":"low-code"},{"id":90,"name":"message-queue"},{"id":42,"name":"mobile-app"},{"id":18,"name":"monitoring"},{"id":36,"name":"networking"},{"id":7,"name":"node-version"},{"id":55,"name":"nosql"},{"id":57,"name":"observability"},{"id":46,"name":"orm"},{"id":52,"name":"os"},{"id":14,"name":"parser"},{"id":74,"name":"react"},{"id":82,"name":"real-time"},{"id":56,"name":"robot"},{"id":65,"name":"runtime"},{"id":32,"name":"sdk"},{"id":71,"name":"search"},{"id":63,"name":"secrets"},{"id":25,"name":"security"},{"id":85,"name":"server"},{"id":86,"name":"serverless"},{"id":70,"name":"storage"},{"id":75,"name":"system-design"},{"id":79,"name":"terminal"},{"id":29,"name":"testing"},{"id":12,"name":"ui"},{"id":50,"name":"ux"},{"id":88,"name":"video"},{"id":20,"name":"web-app"},{"id":35,"name":"web-server"},{"id":43,"name":"webassembly"},{"id":69,"name":"workflow"},{"id":87,"name":"yaml"}]" returns me the "expected json"