Run SWIRL AI Search in a private deployment you control. One click deploys the complete stack - federated search, AI answers with citations, semantic caching, and document intelligence - into your own Railway workspace, on your own subscription. Your sources, your keys, your data: nothing is shared.
Usually within one business day, to the work address you enter above - a block of JSON you paste at deploy time.
Sign in to Railway, paste your license, and deploy. Add an Anthropic API key for fast AI answers with Claude - or connect OpenAI, Azure OpenAI, Gemini, and others after deploy.
Log in and search the built-in sources right away, then connect your own - Microsoft 365, Google Workspace, databases, and enterprise systems.
Your SWIRL Enterprise license, plus Railway infrastructure billed to your own subscription: the Pro plan is $20/month per workspace (unlimited seats, includes $20 of usage), then usage-based compute at roughly $10 per GB of memory and $20 per vCPU per month, billed per second for what the stack actually uses. A small-team deployment typically runs in the low hundreds of dollars per month all-in, and you can watch live usage in the Railway dashboard.
Railway has no instance types - every service scales vertically on demand, up to 1,000 vCPU and 1 TB of memory per service on Pro, and you pay only for what's used. Horizontal scaling with up to 42 replicas is available for high-traffic deployments.