Skip to content

Glossary

A single canonical place for product terms. If something in a guide or reference page uses a word here, this is what it means.

A PlaidCloud login. An account belongs to one or more organizations and is granted access to specific workspaces within those organizations.

A project-scoped chat interface inside PlaidCloud for asking questions about data, generating expressions, and operating on projects in natural language. See AI Assistant guide.

A workflow operation that spreads values from one set of rows (“source”) to another (“target”) based on driver data and rules. Used for transfer pricing, activity-based costing, IT chargeback, profitability analysis, and similar cost-distribution problems. See Allocations.

A saved configuration that lets PlaidCloud reach an external system — a database, cloud storage account, ERP, REST API. Connections are reused by workflow steps so credentials live in one place. See Connections (guide) · Connectors (reference).

A published, interactive view of project data built from tables and views. See Dashboards.

A hierarchy — used for slicing and aggregating data. Cost centers, products, geography, and time periods are common dimensions. Allocations use dimensions to decide what to allocate to what. See Dimensions.

In an allocation, the data that determines how values are spread across targets. Example: if you’re allocating IT cost to business units by headcount, the headcount table is the driver.

PlaidCloud’s underlying analytic data store. The expression library is versioned by Lakehouse engine: Lakehouse v1 (Databend-based) and Lakehouse v2 (StarRocks-based). See Expressions reference.

Model Context Protocol — the standard PlaidCloud uses to integrate with AI coding agents (Claude Code, Cursor, Copilot, etc.). See AI coding agents.

A user account granted access to a workspace. Members are assigned to security groups within the workspace, which control what they can do. See Access management.

The top-level billing and identity boundary in PlaidCloud. An organization contains one or more workspaces and one or more member accounts.

A custom interactive application built on PlaidCloud data, typically using the Panel framework. See Panel apps.

PlaidCloud’s on-premises connector daemon — runs as a Windows service, Unix/Linux/Mac daemon, or container service to bridge PlaidCloud workflows to local data sources. See PlaidLink reference.

The PlaidCloud Excel add-in for retrieving and refreshing data from project tables inside Excel. See PlaidXL reference.

A unit of work inside a workspace. Each project owns its own tables, workflows, dimensions, and audit history. Projects are isolated from each other — they don’t share state. See Projects.

Make a project table available for external consumption (dashboards, BI tools, downstream systems). See Publishing data.

A bundle of permissions inside a workspace. Members are assigned to security groups to grant them specific capabilities (read tables, edit workflows, manage members, etc.). See Security groups.

A single operation inside a workflow — import a CSV, join two tables, send a notification, run an allocation. Steps come in categories (import, export, tables, allocation, dimension, document, notifications, etc.). See Workflow steps reference.

Structured row-and-column data inside a project. Tables come from imports, transformations, or external sources.

A saved query result layered on top of one or more tables. Behaves like a table for downstream consumers but the underlying data is re-evaluated on read.

A sequence of steps that operates on tables. The unit of automation in PlaidCloud. Workflows can run sequentially, in parallel, conditionally, or in loops. See Workflows (guide) · Workflow steps (reference).

An isolated environment inside an organization where actual work happens. Teams typically use one workspace per environment (dev, staging, prod) or per business unit. Members are granted access at the workspace level. See Access management.