Glossary
Common terms used across the Pluggedspace platform.
| Term | Definition |
|---|---|
| Agent | An AI assistant that performs tasks using tools, skills, and company knowledge under your supervision. |
| Agent Fleet | The set of pre-built and custom agents running in your workspace (e.g., Monica, Marketer, Business Finder). |
| Approval Inbox | The Human-in-the-Loop queue where high-risk agent actions wait for human sign-off. |
| Atlas | The Pluggedspace intelligence and data suite for ingestion, analytics, and predictive models. |
| Autonomy Mode | How much control an agent has: Manual, Hybrid, or Autonomous. |
| API Key | A workspace-scoped secret used to authenticate programmatic requests. |
| Console V2 | The Pluggedspace Agentic Operating System research prototype, preserved as a technical architecture and design archive. |
| Custom Agent | An agent built with the visual builder for your exact business workflows. |
| Data Pipeline | A scheduled workflow that cleans, transforms, and enriches datasets automatically. |
| Domain | A data category assigned to a dataset in Atlas (e.g., Sales, Marketing, Operations) — or a verified website a widget is locked to. |
| Domain Locking | Security feature that restricts chat widgets to verified company domains. |
| HITL | Human-in-the-Loop — involving human approval in agent workflows for high-risk actions. |
| Integrations | Connections to third-party services (Stripe, Slack, Email, CRM, GitHub, …) via API keys or OAuth. |
| Marketplace | A catalog of community agent packages and workflow templates you can install or publish. |
| Memory Bank | The workspace's semantic knowledge base (RAG) that agents reference in responses. |
| NLP Querying | Asking questions in natural language instead of SQL. |
| Policy Guardrail | A rule limiting agent behavior (spending limits, rate limits, time windows, approval requirements). |
| RAG | Retrieval-Augmented Generation — fetching relevant company documents to ground agent responses. |
| Role-Based Access Control (RBAC) | Permissions model: Workspace Owner, Admin, Analyst, and Viewer. |
| Skills | Reusable multi-step capabilities (e.g., WebsiteAuditor, SEOAnalyzer, InvoiceExtractor) attachable to agents. |
| SQL Studio | The fast in-memory query editor for running SQL against Atlas datasets. |
| Tenant / Workspace | Your organization's isolated environment — data, memory, and configuration never leak across tenants. |
| Tools | Native capabilities agents use: Email, Slack, WhatsApp, web scraping, databases, payments, and more. |
| Trigger Keywords | Words that activate a custom agent automatically. |