Concept
Glossary
Every Lemma term, defined in one line, A to Z.
Pod
A self-contained workspace for one team or process — its agents, data, flows, and permissions live together.
Agent
An AI worker with a role, instructions, and scoped access to your tables, files, and apps.
Workflow
A repeatable process: steps of agents, functions, decisions, and human approvals that run in order.
Table
Typed business data the pod reads and writes — leads, tickets, tasks — with per-row security.
File
Documents the pod can search, read, and cite in its answers.
Surface
A channel where work reaches the pod: Slack, Gmail, WhatsApp, Telegram, Teams, Outlook.
App
A custom app where your team and the pod’s agents work together — built on this pod’s data, deployed at its own URL.
Kit
A whole pod setup compressed into something shareable — install it like a plugin and it builds out apps, agents, tables, workflows, and schedules.
Function
Deterministic code for steps that shouldn’t be AI judgment: validation, math, formatting.
Schedule
A timer that runs an agent or flow automatically — daily digests, hourly syncs.
Connector
An authenticated connection to a third-party app that agents and flows can act on.
Conversation
A thread between you (or a surface) and this pod’s agents, with every tool call visible.
Sharing
Who can see and use a resource: private, pod members, or the whole organization.
Access scope
The tables, folders, tools, and apps this worker is allowed to read and act on — nothing else.
Approval
A pause in a flow where a named person decides before work continues.
Runtime
The model and harness an agent runs on — pick per agent.
Variables
The typed inputs an agent expects and the structured output it returns — its contract with the rest of the pod.