# BAML — Schema-First Prompting for Agents

> BAML treats prompts as typed functions for reliable outputs. Install `baml-py`, define `.baml` functions, then call generated clients from Python.

## Install

Save as a script file and run:

# BAML — Schema-First Prompting for Agents

> BAML treats prompts as typed functions for reliable outputs. Install `baml-py`, define `.baml` functions, then call generated clients from Python.

## Quick Use

1. Install:
   ```bash
   pip install baml-py
   ```
2. Run:
   ```bash
   python -c "import baml_py; print('baml-py installed')"
   ```
3. Verify:
   - Open the repo README and confirm you can define a BAML function and call generated `baml_client` from Python


---

## Intro

BAML treats prompts as typed functions for reliable outputs. Install `baml-py`, define `.baml` functions, then call generated clients from Python.

- **Best for:** teams iterating on prompts daily who want typesafety, streaming, and retry-friendly workflows without rewriting the whole app
- **Works with:** BAML `.baml` files, Python via `baml-py`, and multi-language clients (per repo)
- **Setup time:** 15 minutes


### Quantitative Notes

- PyPI package: `baml-py` (verified via pip index)
- Setup time ~15 minutes
- GitHub stars (verified): see Source & Thanks


---

## Practical Notes

Adopt BAML when your bottleneck is *iteration speed* and *output stability*. Model your prompt as a function: inputs are typed, outputs are typed, and the prompt body is versionable. Keep a small library of BAML functions (extract, classify, route, summarize) and test them in your IDE before wiring them into agents. Over time, you get a maintainable prompt surface instead of a pile of ad-hoc strings.

**Safety note:** Keep function signatures small; overly generic types and huge prompts slow iteration and reduce reliability.

### FAQ

**Q: Do I need to write my whole app in BAML?**
A: No. The repo emphasizes using BAML for prompts/functions and wiring it into your existing app.

**Q: What makes it different from plain prompting?**
A: You define a function signature with types and let the tool enforce output structure, retries, and streaming-friendly interfaces.

**Q: Is it Python-only?**
A: No. The repo lists Python, TypeScript, Ruby, Go, and more client options via generated clients or REST patterns.

---

## Source & Thanks

> GitHub: https://github.com/BoundaryML/baml
> Owner avatar: https://avatars.githubusercontent.com/u/124114301?v=4
> License (SPDX): Apache-2.0
> GitHub stars (verified via `api.github.com/repos/BoundaryML/baml`): 8,206


---

<!-- ZH -->

# BAML——以 Schema 为中心的 Prompt 工程

> BAML 把 prompt 当作「带类型的函数」来写，提升输出稳定性（即使模型不支持 tool-calling）。安装 `baml-py`，定义 `.baml` 函数，再从 Python 调用。

## 快速使用

1. 安装：
   ```bash
   pip install baml-py
   ```
2. 运行：
   ```bash
   python -c "import baml_py; print('baml-py installed')"
   ```
3. 验证：
   - Open the repo README and confirm you can define a BAML function and call generated `baml_client` from Python


---

## 简介

BAML 把 prompt 当作「带类型的函数」来写，提升输出稳定性（即使模型不支持 tool-calling）。安装 `baml-py`，定义 `.baml` 函数，再从 Python 调用。

- **适合谁（Best for）:** 每天迭代 prompt 的团队，希望引入类型安全、流式与可重试工作流，但不想重写整个应用
- **兼容工具（Works with）:** BAML `.baml` 文件、Python 的 `baml-py`，以及多语言 client（仓库说明）
- **安装时间（Setup time）:** 15 分钟


### 量化信息

- PyPI 包：`baml-py`（已用 pip index 核验）
- 装机约 15 分钟
- GitHub stars（已核验）：见「来源与感谢」


---

## 实战要点

当你的瓶颈是「迭代速度」与「输出稳定性」时，BAML 很适合：把 prompt 建模成函数，输入/输出都有类型，prompt 本身可版本化。把常用能力沉淀成一组 BAML 函数（抽取/分类/路由/总结），先在 IDE 里快速测试，再接入 agent。长期看，你会得到可维护的 prompt 工程面，而不是一堆零散字符串。

**安全提示：** 函数签名要小而清晰：类型过于泛化、prompt 过大，会拖慢迭代并降低稳定性。

### FAQ

**Q: 必须用 BAML 重写整个应用吗？**
A: 不需要。仓库强调只把 BAML 用在 prompt/function 层，再接入你现有的业务代码。

**Q: 和普通 prompt 有什么不同？**
A: 你先定义函数签名与类型，让工具帮助你把输出结构、重试与流式接口工程化，而不是靠临时提示词凑结果。

**Q: 只能 Python 吗？**
A: 不是。仓库提到 Python、TypeScript、Ruby、Go 等多语言调用方式（生成 client 或 REST）。

---

## 来源与感谢

> GitHub：https://github.com/BoundaryML/baml
> Owner avatar：https://avatars.githubusercontent.com/u/124114301?v=4
> 许可证（SPDX）：Apache-2.0
> GitHub stars（已通过 `api.github.com/repos/BoundaryML/baml` 核验）：8,206


---
Source: https://tokrepo.com/en/workflows/baml-schema-first-prompting-for-agents
Author: Agent Toolkit