kyegomez/swarms

What is it?

Key features

• 01Sequential, concurrent, graph (DAG) and hierarchical multi-agent workflows.
• 02SwarmRouter provides one interface to run different swarm types.
• 03Agents connect to MCP servers and can also be served as MCP servers.
• 04Supports agent skills defined in SKILL.md files.
• 05Automatically generates agents and workflows from a task description with AutoSwarmBuilder.

Use cases

• Building a research pipeline where one agent gathers material and another drafts the output.
• Running several specialist agents in parallel on the same analysis task and merging their results.
• Exposing an internal agent to other MCP-compatible hosts behind an API key.
• Prototyping a multi-agent system quickly from a plain-language task description.

Quick start

1. 1

   Install the package with pip

   pip3 install -U swarms
2. 2

   Set your model provider API key in the environment

3. 3

   Run the first agent example in Python

   python agent.py

Things to know

• !Examples use hosted LLM APIs, so you need a paid provider API key to run them.
• !The README describes many architectures, so the API surface is large and may take time to learn.

Topics

FAQ

What is swarms?

Python framework for building and orchestrating multi-agent systems with sequential, concurrent, graph-based and hierarchical workflows. Swarms is a Python framework for building systems where multiple LLM-powered agents work together on a task. Each Agent wraps a model with a system prompt, tools, and optional memory, and the framework provides several ready-made orchestration patterns to coordinate them, including SequentialWorkflow, ConcurrentWorkflow, AgentRearrange, GraphWorkflow, MixtureOfAgents, HierarchicalSwarm, and a SwarmRouter that selects among them.

Is swarms free to use? What license does it have?

swarms is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.

Is swarms still maintained?

Actively maintained. The last commit to swarms was on Oct 19, 2026.

What can I use swarms for?

Building a research pipeline where one agent gathers material and another drafts the output. Running several specialist agents in parallel on the same analysis task and merging their results. Exposing an internal agent to other MCP-compatible hosts behind an API key. Prototyping a multi-agent system quickly from a plain-language task description.

How do I install swarms?

Install the package with pip: pip3 install -U swarms Set your model provider API key in the environment Run the first agent example in Python: python agent.py

Where can I get swarms?

The source code and installation instructions are on GitHub: https://github.com/kyegomez/swarms