Apache Burr：构建可靠的人工智能代理和应用程序
阅读原文· burr.apache.orgApache Burr 是一个用于构建可靠 AI 智能体和应用程序的框架，已在 Apache 基金会下发布。该项目提供工具和抽象，帮助开发者设计、开发和部署可信任的智能体应用，强调可靠性、可观测性和生产级部署能力。
Build reliable AI agents and applications
Apache Burr (Incubating) makes it easy to develop applications that make decisions, from simple chatbots to complex multi-agent systems. Pure Python, no magic.
Simple, powerful Python API
Build anything from chatbots to multi-agent systems with a clean, composable interface.
from burr.core import action, State, ApplicationBuilder
@action(reads=["messages"], writes=["messages"])
def chat(state: State, llm_client) -> State:
response = llm_client.chat(state["messages"])
return state.update(
messages=[*state["messages"], response]
)
app = (
ApplicationBuilder()
.with_actions(chat)
.with_transitions(("chat", "chat"))
.with_state(messages=[])
.with_tracker("local")
.build()
)
app.run(halt_after=["chat"], inputs={"llm_client": client})Everything you need to build AI applications
Burr provides the building blocks for reliable, observable, and testable AI-powered applications.
Simple Python API
Define your application as a set of actions and transitions. No DSL, no YAML — just Python functions and decorators.
Built-in Observability
The Burr UI lets you monitor, debug, and trace every step of your application in real time. See state changes as they happen.
Persistence & State Management
Automatically persist state to disk, databases, or custom backends. Resume applications from where they left off.
Human-in-the-Loop
Pause execution and wait for human input at any step. Perfect for approval workflows and interactive agents.
Branching & Parallelism
Run actions in parallel, fan out / fan in, and build complex DAGs. Compose sub-applications for modular design.
Testing & Replay
Replay past runs, unit test individual actions, and validate state transitions. Build confidence in your AI systems.
Works with your stack
Burr integrates with the tools and frameworks you already use. No lock-in, no wrappers.
OpenAI
LLMAnthropic
LLMLangChain
FrameworkHamilton
FrameworkStreamlit
UIOpenAI
LLMAnthropic
LLMLangChain
FrameworkHamilton
FrameworkStreamlit
UIOpenAI
LLMAnthropic
LLMLangChain
FrameworkHamilton
FrameworkStreamlit
UIOpenAI
LLMAnthropic
LLMLangChain
FrameworkHamilton
FrameworkStreamlit
UIFastAPI
ServingHaystack
FrameworkInstructor
LLMPydantic
ValidationPostgreSQL
StorageFastAPI
ServingHaystack
FrameworkInstructor
LLMPydantic
ValidationPostgreSQL
StorageFastAPI
ServingHaystack
FrameworkInstructor
LLMPydantic
ValidationPostgreSQL
StorageFastAPI
ServingHaystack
FrameworkInstructor
LLMPydantic
ValidationPostgreSQL
StorageTrusted by engineers worldwide
See what developers and teams are saying about Burr.
“After evaluating several other obfuscating LLM frameworks, their elegant yet comprehensive state management solution proved to be the powerful answer to rolling out robots driven by AI decision making.”
“Using Burr is a no-brainer if you want to build a modular AI application. It is so easy to build with and I especially love their UI which makes debugging a piece of cake. And the always ready to help team is the cherry on top.”
“I just came across Burr and I'm like WOW, this seems like you guys predicted this exact need when building this. No weird esoteric concepts just because it's AI.”
“Burr's state management part is really helpful for creating state snapshots and build debugging, replaying and even building evaluation cases around that.”
“I have been using Burr over the past few months, and compared to many agentic LLM platforms out there (e.g. LangChain, CrewAi, AutoGen, Agency Swarm, etc), Burr provides a more robust framework for designing complex behaviors.”
“Moving from LangChain to Burr was a game-changer! It took me just a few hours to get started with Burr, compared to the days and weeks I spent trying to navigate LangChain. I pitched Burr to my teammates, and we pivoted our entire codebase to it.”
“Of course, you can use it [LangChain], but whether it's really production-ready and improves the time from code-to-prod, we've been doing LLM apps for two years, and the answer is no. Honestly, take a look at Burr. Thank me later.”
“After evaluating several other obfuscating LLM frameworks, their elegant yet comprehensive state management solution proved to be the powerful answer to rolling out robots driven by AI decision making.”
“Using Burr is a no-brainer if you want to build a modular AI application. It is so easy to build with and I especially love their UI which makes debugging a piece of cake. And the always ready to help team is the cherry on top.”
“I just came across Burr and I'm like WOW, this seems like you guys predicted this exact need when building this. No weird esoteric concepts just because it's AI.”
“Burr's state management part is really helpful for creating state snapshots and build debugging, replaying and even building evaluation cases around that.”
“I have been using Burr over the past few months, and compared to many agentic LLM platforms out there (e.g. LangChain, CrewAi, AutoGen, Agency Swarm, etc), Burr provides a more robust framework for designing complex behaviors.”
“Moving from LangChain to Burr was a game-changer! It took me just a few hours to get started with Burr, compared to the days and weeks I spent trying to navigate LangChain. I pitched Burr to my teammates, and we pivoted our entire codebase to it.”
“Of course, you can use it [LangChain], but whether it's really production-ready and improves the time from code-to-prod, we've been doing LLM apps for two years, and the answer is no. Honestly, take a look at Burr. Thank me later.”
“After evaluating several other obfuscating LLM frameworks, their elegant yet comprehensive state management solution proved to be the powerful answer to rolling out robots driven by AI decision making.”
“Using Burr is a no-brainer if you want to build a modular AI application. It is so easy to build with and I especially love their UI which makes debugging a piece of cake. And the always ready to help team is the cherry on top.”
“I just came across Burr and I'm like WOW, this seems like you guys predicted this exact need when building this. No weird esoteric concepts just because it's AI.”
“Burr's state management part is really helpful for creating state snapshots and build debugging, replaying and even building evaluation cases around that.”
“I have been using Burr over the past few months, and compared to many agentic LLM platforms out there (e.g. LangChain, CrewAi, AutoGen, Agency Swarm, etc), Burr provides a more robust framework for designing complex behaviors.”
“Moving from LangChain to Burr was a game-changer! It took me just a few hours to get started with Burr, compared to the days and weeks I spent trying to navigate LangChain. I pitched Burr to my teammates, and we pivoted our entire codebase to it.”
“Of course, you can use it [LangChain], but whether it's really production-ready and improves the time from code-to-prod, we've been doing LLM apps for two years, and the answer is no. Honestly, take a look at Burr. Thank me later.”
“After evaluating several other obfuscating LLM frameworks, their elegant yet comprehensive state management solution proved to be the powerful answer to rolling out robots driven by AI decision making.”
“Using Burr is a no-brainer if you want to build a modular AI application. It is so easy to build with and I especially love their UI which makes debugging a piece of cake. And the always ready to help team is the cherry on top.”
“I just came across Burr and I'm like WOW, this seems like you guys predicted this exact need when building this. No weird esoteric concepts just because it's AI.”
“Burr's state management part is really helpful for creating state snapshots and build debugging, replaying and even building evaluation cases around that.”
“I have been using Burr over the past few months, and compared to many agentic LLM platforms out there (e.g. LangChain, CrewAi, AutoGen, Agency Swarm, etc), Burr provides a more robust framework for designing complex behaviors.”
“Moving from LangChain to Burr was a game-changer! It took me just a few hours to get started with Burr, compared to the days and weeks I spent trying to navigate LangChain. I pitched Burr to my teammates, and we pivoted our entire codebase to it.”
“Of course, you can use it [LangChain], but whether it's really production-ready and improves the time from code-to-prod, we've been doing LLM apps for two years, and the answer is no. Honestly, take a look at Burr. Thank me later.”
Join the community
Get help, share your projects, and contribute to the future of Burr.