Core concepts

A short glossary of the terms you'll meet across the product. Each links to the page that covers it in depth.

Workspace

The account that owns everything you build: skills, playbooks, credentials, variables, and credits. Today a workspace belongs to one account; team workspaces are not available yet.

Skill

A packaged connection to one application's API — a manifest of operations (each one a concrete API call, marked read or write) plus documentation. Skills come from the gallery, where every skill's read operations were live-tested against the real API before it shipped. Once added, a skill shows up under My Skills with a status badge: Ready, Partial, Generating, or Failed.

Credential

The API token a skill needs to reach its application, collected through a secure card and stored write-only in an encrypted vault. Credentials are per API domain, shared by every skill in your workspace that talks to that domain. See Connecting your app.

Variable

A per-skill default value for the parameters its operations take (a project key, an account id, a region). Defaults feed both test chat and playbook runs, and can be overridden per chat session. Managed on the skill page's Variables card.

Test chat

The Test against live API panel on every skill page: a conversation that actually executes the skill's operations and shows each call as expandable evidence. Read-only unless you enable write calls for the session. See Test chat.

Playbook

A standing order: instructions in plain words, one or more mounted skills, and a schedule. Playbooks run unattended and are always read-only. A playbook is Active or Paused; your plan caps how many can be active at once. See Playbooks.

Run

One execution of a playbook — scheduled, or started by you as a Test run. A run records its output and one evidence block per API call, and lands in the playbook's history and your runs feed. See Runs & history.

Runs feed

The stream on your home page where playbook activity lands, topped by a once-daily recap that summarises what ran and what it found.

Credit

The single unit of metered usage. Exactly two things spend credits: test-chat turns and playbook runs. Everything else — setup conversations, adding skills, connecting apps, drafting playbooks — is free. See Credits & plans.

Weight class

What sets an action's credit price. Every skill is Light (up to 20 operations), Standard (21–50), or Heavy (51 or more); heavier skills carry more operations into every model call, so they cost more per turn and per run. A playbook is weighed by all its mounted skills together.

Trial and Pro

Every new workspace starts on the Trial: a one-time grant of 100 credits with no card and no expiry clock. Pro is the paid plan — monthly credits that reset on your billing date, more active playbooks, and hourly schedules. See Trial & Pro.

Stopped

The workspace state at zero credits: chat and runs pause, everything you built and every past run stays readable, and scheduled playbooks resume by themselves once credits are back. Nothing is ever disabled or deleted.