Get started with Swirls

Swirls is the managed platform for systems defined in .swirls files. Compose agents, workflows, data, connections, access rules, and apps. Validate locally, deploy to Cloud, and operate everything as one project.

GET UP AND RUNNING

Your first deployment

Install the CLI, create a project, validate your files, and inspect the deployed system in Cloud.

Start the quickstart →

UNDERSTAND SWIRLS

From source to running system

Learn how projects, deployment snapshots, managed resources, and hosted execution fit together.

Explore the platform →

01 Define02 Validate03 Deploy04 Operate

The platform

Understand what Swirls manages, where your code runs, and how to operate a deployed project.

Projects & deployments

Source files, immutable snapshots, active deployments, Git integration, and rollback.

Deployment guide →

Managed infrastructure

Execution workers, isolated code, agent sandboxes, persistent storage, credentials, and networking.

Infrastructure overview →

Execution & operations

How work starts, checkpoints, waits, and recovers. Test with playbooks and inspect runs through traces.

How execution works →

The primitives

Primitives are the building blocks of a .swirls project. Each declares part of the system; the platform connects and runs them together.

Agents

Model-driven reasoning with tools, skills, persistent sessions, and chat channels.

Workflows

Explicit steps and data flow, started by triggers, with branching, waits, and human reviews.

Memory

Structured records, views, managed databases, existing Postgres, and persistent files.

Connections

Secrets, credential profiles, OAuth connections, typed API actions, and private networks.

Access

Identity-derived roles and policies that control who can invoke agents and workflows.

Interfaces

Hosted apps that bring pages, data, chat, and workflow reviews into a workspace.

See how the primitives fit together →

Keep building