Currently in alpha

The AI platform for operational builders.

MSPs, forward-deployed engineers, and GTM engineers use Swirls to turn operational knowledge into AI systems they can ship, run, and grow for every customer.

Start building
Quotes sentTickets closedFollow-ups handled
A repeatable delivery model

Build the expertise once. Put it to work everywhere.

Swirls turns the way your team delivers outcomes into a system you can adapt, operate, and improve across customers.
01

Capture the operating model

Turn the process, judgment, and systems behind the work into one clear, reviewable AI system.

02

Adapt it for every customer

Start from a proven foundation, then tailor the connections, rules, and experience for each engagement.

03

Operate, prove, and expand

See what ran, control what it can touch, and grow from one successful outcome into the next.

Operational builders at work

BrewTech built an agency, not another demo.

BrewTech uses Swirls to operate specialists under one orchestrator—including a live quoting system for a manufacturing customer.
Customer story
Swirls lets me build an entire agency on top of it. I take it and run.
Ben BrewerVP of Engineering, BrewTech IT
Read the BrewTech story
Business model
AI agent agency
Live outcome
Manufacturing quoting, end to end
Operating model
Specialists under one orchestrator
Trust is part of the delivery

Built to earn the next engagement.

Operational builders are accountable for what their systems do. Swirls makes isolation, permissions, and auditability part of the work from day one.

Keep customers isolated

Each customer's connections and work stay in their own boundary, from the first run onward.

Control what agents can do

Set clear roles and permissions for every agent, tool, and team. Start narrow, then expand deliberately.

Prove what happened

Review the system alongside every change and trace what it did when customers or auditors ask.

Under the hood

One declaration. A complete platform.

When your technical team wants the details, the whole system is there: readable source in, durable and governed runtime out.
01 · Declare

Describe the system.

Agents and deterministic tools live together in readable, reviewable files.

acme · concierge
agent.swirls
concierge › agent.swirls
secret provider_keys {
  label: "Provider keys"
  vars: [OPENROUTER_API_KEY]
}

agent concierge {
  label: "Concierge"
  secrets: provider_keys
  provider: openrouter
  model: "openai/gpt-4o-mini"
  // See tools.swirls for workflow definitions.
  tools: [search_orders]
  system: @ts {
    return "Help customers with orders, returns, and shipping updates."
  }
}
terminal
~/acme $swirls doctor
✓ 1 agent · 2 workflows · 2 triggers

Turn the next operation into an AI outcome.

Start with one task. Build the system around it. Run it for one customer, then every customer.