Open Swarm
by matthewhand
Open Swarm is a versatile, modular framework for building intelligent, multi-agent systems. It's an actively maintained extension of the OpenAI Swarm framework, modified to support stateless RESTful operations and a plugin system for custom extensions that enhance agentic workflows.
Last updated: N/A
What is Open Swarm?
Open Swarm is a versatile and modular framework designed for building intelligent, multi-agent systems. It's a fork of the OpenAI Swarm framework with modifications for stateless RESTful operations and a plugin system.
How to use Open Swarm?
To use Open Swarm, install the package using pip, configure an LLM provider, optionally configure an MCP server, add a blueprint using the Swarm CLI, and then run the blueprint from the CLI. The framework also exposes a REST API for external integration.
Key features of Open Swarm
Multi-Agent Orchestration
Blueprint-Driven Architecture
Optional MCP Integration
CLI & REST Interface
OpenAI API Compatibility
Configurable LLMs
Use cases of Open Swarm
Building personal assistants
Creating system administration tools
Orchestrating multi-agent conversations
Integrating external tools and databases
FAQ from Open Swarm
What is a Blueprint?
What is a Blueprint?
A Blueprint is a Python module that wraps agent logic, tool connections, and environment/configuration settings.
How do I configure an LLM provider?
How do I configure an LLM provider?
When you run a blueprint for the first time, Open Swarm checks for a configuration file at ~/.swarm/swarm_config.json. If the file is missing, it will automatically create a default configuration.
What is an MCP server?
What is an MCP server?
MCP servers integrate external tools (e.g., databases, web search, filesystems) into the agent workflows.
Is Open Swarm compatible with the OpenAI API?
Is Open Swarm compatible with the OpenAI API?
Yes, Open Swarm exposes an endpoint at /v1/chat/completions that is functionally similar to the OpenAI Chat Completions API.
How do I add an example blueprint?
How do I add an example blueprint?
Add an example blueprint by running: swarm-cli add /path/to/your/blueprint.py --name example