Burr

Burr: Apache Burr is an open-source Python framework for building, observing, and testing AI agents and applications that make decisions.

Apache Burr (Incubating) enables developers to create AI-powered applications, ranging from chatbots to multi-agent systems, using standard Python functions and decorators. The framework emphasizes modular design, allowing users to define applications as sets of actions and transitions without proprietary DSLs or YAML configurations. It includes built-in observability tools for real-time monitoring, debugging, and tracing of state changes. Developers can manage state persistence across various backends, implement human-in-the-loop workflows, and execute parallel or branched logic. The framework integrates with existing stacks, including tools like OpenAI, Anthropic, LangChain, Hamilton, Streamlit, FastAPI, Haystack, Instructor, Pydantic, and PostgreSQL.

Burr at a glance

Best for: Software engineer
Runs on: self-hosted, open source
Pricing: Free
Released: January 2024
Category: AI coding tool
Website: burr.apache.org

Key features

Python-native API without custom DSLs or YAML
Real-time monitoring and debugging UI
Automated state persistence to disk or databases
Human-in-the-loop pause and approval workflows
Parallel action execution and branching support
Replay capabilities for testing and evaluation

Use it for

Building AI agents with complex decision logic
Creating interactive, human-in-the-loop chat interfaces
Developing modular, observable multi-agent systems
Testing and debugging AI state transitions

Alternatives to Burr

OpenSquilla — Maximizes AI intelligence density while minimizing token costs through smart routing.
Beeuni — Agents retain context and improve skills over time through a controlled review process.
BeeAI Framework — Provides constraint enforcement for predictable agent behavior.
Atomic Agents — Build predictable, modular AI applications with consistent, reliable outputs.
Marvin — Provides simple interfaces, smart defaults, and structured data extraction.
ARTEX — Runs autonomous multi-agent penetration testing with a dual-graph architecture.
AG-UI — Standardizes agent-user interactions and generative UI for applications.
Kiln — Combines a user-friendly desktop app with an open-source Python library.
OpenAPPA — Tracks data flows algebraically to prevent prompt injection without breaking agents.
OpenSandbox — Provides isolated environments with multi-language SDKs and lifecycle management.
NemoClaw — Provides managed security, networking, and lifecycle operations for AI agents.
Cairn — Solves open-ended problems using a blackboard architecture and OODA loops.

All Burr alternatives