Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/UnkleFunk/HouseMusicSwarm-/llms.txt

Use this file to discover all available pages before exploring further.

By the end of this guide you’ll have OpenSwarm running in your terminal, authenticated with your chosen AI provider, and ready to accept your first prompt. The recommended path takes under 60 seconds: one npm install command launches a setup wizard that handles Python dependencies, provider authentication, and optional integrations automatically.
1

Install OpenSwarm

OpenSwarm is distributed as an npm package. Install it globally with:
npm install -g @vrsen/openswarm
Requirements: Node.js 18+ must be installed on your machine. Python 3.12+ is detected and installed automatically — you do not need to set up a Python environment yourself.After installation completes, run:
openswarm
This launches the interactive setup wizard. If Python dependencies are missing, they are installed silently on first run using uv.
2

Run the setup wizard

The setup wizard walks you through three steps:Step 1 — Choose your AI provider. Select one of the three supported providers:
  • OpenAI — uses gpt-5.2 by default; required for Sora video generation
  • Anthropic — uses litellm/claude-sonnet-4-6 by default; recommended for best slide HTML output
  • Google Gemini — uses litellm/gemini/gemini-3-flash by default; required for Gemini image generation and Veo video
Step 2 — Enter your API key. The wizard prompts for your provider’s API key (input is masked). Your key is written directly to .env — it never leaves your machine.Step 3 — Enable optional add-ons. Select any combination of:
Add-onWhat it unlocks
Web SearchWeb, Scholar & product search for all agents (SEARCH_API_KEY)
Composio10,000+ integrations — Gmail, Slack, GitHub, HubSpot (COMPOSIO_API_KEY + COMPOSIO_USER_ID)
Google GeminiImage generation, editing, and Veo video (GOOGLE_API_KEY)
fal.aiSeedance 1.5 Pro video, editing, background removal (FAL_KEY)
Stock photosPexels / Pixabay / Unsplash image search for Slides Agent
Once the wizard completes, OpenSwarm starts automatically and drops you into the interactive terminal UI.
All tools gracefully degrade when optional API keys are missing. You’ll receive a clear message explaining what to add rather than a silent failure — so it’s safe to start without every key and add them later.
3

Send your first prompt

Type a prompt directly into the terminal UI. Here are a few to get you started:
Create a complete investor pitch for my startup

Research my top 5 competitors and write 3 SEO-optimized blog posts

Analyze this CSV and create a quarterly report with charts

Generate a product launch video with animations
The Orchestrator receives your message, breaks it into parallel workstreams, and routes each piece to the right specialist. Completed files are saved locally and their paths are printed in the response.
For a full reference of every environment variable — including all provider keys, model selection, and optional integration keys — see the Configuration page.

​
Alternative installation methods

git clone https://github.com/VRSEN/openswarm.git
cd openswarm
python swarm.py

​
Local development

Clone the repo and run python swarm.py directly to launch the Agency Swarm terminal UI with full show_reasoning=True output. This is the best option if you want to customize agents, add tools, or extend the swarm for your own use case. Dependencies are bootstrapped automatically via run_utils._bootstrap() on first run.

​
Docker

Copy .env.example to .env, add your keys, then run docker-compose up --build. Docker is the easiest way to deploy OpenSwarm on a server or share a reproducible environment with a team.

​
API server

python server.py starts a FastAPI server on localhost:8080. This exposes OpenSwarm’s agent orchestration over HTTP, making it easy to integrate the swarm into existing applications or trigger agents from CI/CD pipelines.

Build docs developers (and LLMs) love

Get started for freeTalk to us