Laddr is a Python framework for building scalable multi-agent systems, functioning as a microservices architecture for AI agents. It offers built-in message queues, observability, and horizontal scalability, enabling agents to communicate, delegate tasks, and execute work in parallel, streamlining the development of complex AI workflows.

How It Works

Laddr supports dynamic Coordinator-Orchestrator workflows and pre-defined Sequential Deterministic pipelines. Its core architecture utilizes Redis Streams for reliable, distributed messaging and automatic load balancing across agent workers. Comprehensive observability is achieved through automatic tracing to PostgreSQL, while large data payloads are managed via MinIO/S3 artifact storage, optimizing message bus performance and handling large documents.

Quick Start & Requirements

• Installation: pip install laddr
• Project Setup: laddr init <my-agent-system>, then cd <my-agent-system>
• Configuration: Edit .env for API keys (Serper, LLMs) or Ollama.
• Prerequisites: Python 3.10+, Redis (7.0+), PostgreSQL (15+), MinIO/S3 (optional), Docker (optional).
• Run: laddr run dev starts the API (http://localhost:8000), dashboard (http://localhost:5173), and supporting services.
• Docs: Quick Start, Documentation

Highlighted Details

• Scalability: Achieved via horizontal scaling, multi-worker support, automatic load balancing, and fault tolerance.
• Observability: Integrated tracing to PostgreSQL, real-time metrics, and an interactive dashboard for monitoring.
• LLM Agnostic: Supports Gemini, OpenAI, Anthropic, Groq, Ollama, and local models.
• Production Ready: Features a FastAPI REST API, Docker native support, and queue-based messaging.
• Extensibility: Custom tools via @tool decorator and overrideable system tools for delegation/storage.

Maintenance & Community

The project is hosted on GitHub with an issues tracker. Links to the website and documentation are provided. Specific community channels or maintainer details are not detailed in the README.

Licensing & Compatibility

Licensed under Apache License 2.0, permitting commercial use and closed-source linking.

Limitations & Caveats

The README emphasizes production readiness and does not detail explicit limitations or known bugs. Setup requires configuring multiple external services (Redis, PostgreSQL, MinIO/S3) and API keys, which may present an initial hurdle.