How the Platform Works
The AI Rule Engine has two sides: an authoring side where you design rule sets, and an executing side where those rule sets run. Understanding this split is the key to everything else in these docs.
Two sides: authoring and executing
You do all of your design work — organizations, projects, rule sets, conditions, and actions — on the authoring side. When you are ready to put a rule set to work, you run it on the executing side, inside an environment. The same rule set can run in several environments without being rebuilt.
| Authoring | Executing |
|---|---|
| Create projects, rule sets, rules, conditions, and actions. | Run rule sets against real input, on demand, on a schedule, or via the API. |
| Reuse conditions, action sets, extensions, and templates. | Apply per-environment settings, variables, and storage. |
| Validate and refine your logic before you ship it. | Produce results, files, and logs you can review afterward. |
| Manage members, plan, and credits for your organization. | Meter runs and enforce plan limits as work happens. |
For a closer look at how a design becomes a running workload, see Authoring & Executing.
What you work with
Everything you do happens in the web app. From there you author rule sets, manage your organization and environments, run rule sets and read their logs, and connect integrations. A few building blocks recur throughout:
- Organizations and projects organize your work and govern who can access it. See Organizations and Projects.
- Rule sets, rules, conditions, and actions are the logic itself. Start at Core Concepts.
- Environments are where rule sets run, each with its own settings and history. See Environments.
- Integrations connect rule sets to the outside world — APIs, AI providers, extensions, and AI assistants. See API Access and MCP Server.
A typical flow
- You author a rule set in a project on the authoring side.
- You choose an environment to run it in.
- You run it — manually, on a schedule, or through the API.
- Rules evaluate against the input, matching rules run their actions, and results are produced.
- You review the outcome in Logs & Run History.