Skip to content
APIANT
GuidePlatformv1

Glossary

Alphabetical lookup for every APIANT platform term — concepts, object types, UI labels, and jargon.

A short lookup of APIANT platform terms in user-facing language. Each entry links to the page that explains it in context.

A

Action — A single operation against an external system: creating a record, fetching one, updating one, etc. Claude picks actions based on what your prompt describes. See Actions.

Assembly — The internal building block APIANT uses to package up a connector or part of one. You don't usually interact with assemblies directly; Claude builds and edits them on your behalf. See Understanding APIANT → Four Core Objects.

Automation — A workflow built from a trigger and one or more actions, with optional conditions and loops. The unit you ask Claude to build. See Understanding APIANT.

C

Catalog — The collection of all built-in connectors APIANT ships. Claude searches it when you describe an integration involving a particular vendor.

Connector — APIANT's representation of an external system: how to authenticate, what events it emits, what operations you can run against it. See Connectors.

Credentials — Auth values APIANT stores in the keyvault. Scoped per user per service by default. See Credentials.

D

Draft — An automation or assembly Claude has built but not yet committed. Claude commits the draft as a versioned release once you've had a chance to review.

E

Execution — A single run of an automation. Available to ask Claude about: "why did the last execution fail?".

F

Field mapping — How data moves between automation steps. See Field Mappings.

I

Integration suite — Several automations that work together as a single deliverable, usually for a customer. See Integration Suites.

K

Keyvalue storage — Built-in per-automation key/value storage with retention tiers (temp_, temp180_, temp365_, no prefix). For state that has to persist across runs.

Keyvault — APIANT's encrypted credential store. Holds OAuth tokens, API keys, and similar secrets. See Credentials.

L

Linked account — Multi-location customer setups: a parent account whose child accounts share credentials and coordinate state. See Tenancy → Linked accounts.

M

MCP — Model Context Protocol. The interface the Claude Code plugin uses to talk to APIANT. See MCP Tool Catalog.

P

Pattern — A reusable shape of automation (chat widget, CSV mapping, latches, subroutines). Claude auto-applies the right pattern based on your prompt. See Patterns.

S

Skill — A workflow the plugin runs when you ask for a specific kind of work. Claude picks the right skill from your description; you don't name it. See Skills Index.

Switch Account — Permission that lets Claude cross between tenants for build, deploy, or support work. See Tenancy → Switching tenants.

T

Template — A canonical automation rolled out across many customer accounts in one coordinated step. Used for productized integrations.

Tenant — A separate database within an APIANT system, used for scaling or isolation when an account isn't enough. Most customers are accounts inside the main tenant, not separate tenants. See Tenancy.

Trigger — The event that starts an automation. Webhook, polling, scheduled, or manual. See Triggers.

Two-way sync — Records kept in step across two systems with loop prevention and last-write-wins resolution. See Two-Way Sync.

W

Webhook — A trigger style: the source system pushes events to APIANT in real time, rather than APIANT polling for them. Claude picks webhook triggers when the source supports them.

Related docs

Last updated May 4, 2026