SWIRL Plugin for Claude Code
The SWIRL plugin gives Claude Code guided, expert workflows for the entire SWIRL lifecycle: installing SWIRL, connecting data sources, configuring AI answers, migrating between editions, troubleshooting, and wiring SWIRL's MCP server into Claude so agents can search your enterprise data with your permissions enforced.
It works with both SWIRL Community and SWIRL Enterprise, and is free and open source under the Apache 2.0 license: github.com/swirlai/swirl-claude-plugin.
One-page overview (PDF): Download the SWIRL Plugin for Claude Code datasheet - no registration required.
Installation
With the Claude Code CLI installed, run:
claude plugin marketplace add swirlai/swirl-claude-plugin
claude plugin install swirl@swirl --scope user
Then start a new Claude Code session - sessions that were already open will not see the new commands.
Commands
| Command | What it does |
|---|---|
/swirl:start | Assesses where you are and routes to the right workflow |
/swirl:install | Guided install - Docker Compose or local - verified with a live search |
/swirl:provider | SearchProvider wizard: connect a source using SWIRL's built-in connectors |
/swirl:connector | Develop a custom connector when no built-in connector fits |
/swirl:rag | Connect an LLM (OpenAI, Anthropic, Azure, Ollama, and more) and tune grounded RAG |
/swirl:migrate | Community-to-Enterprise migration, or version upgrades, with verification |
/swirl:mcp | Wire the SWIRL MCP server into Claude so agents search through SWIRL (Enterprise) |
/swirl:troubleshoot | Log-driven diagnosis of a misbehaving deployment |
Evaluate SWIRL in About an Hour
A typical first session, on your own data:
/swirl:install # get the stack running, verified with a live search
/swirl:provider # connect your first two or three sources
/swirl:rag # grounded answers with citations, from your data
Because SWIRL is federated - data stays in its sources, nothing is copied or indexed - there is no ingestion step between "installed" and "searching your own content." That is what makes a one-hour evaluation realistic.
Editions
All commands support both editions and call out the differences where they matter: install paths, RAG configuration (Community uses an OpenAI key via environment variable; Enterprise uses configurable AI Providers), and the MCP server, which is a SWIRL Enterprise feature.
Get Help
Issues and contributions are welcome on the GitHub repository. For SWIRL itself, see the Troubleshooting Guide or email support@swirlaiconnect.com.