Glossary
The terms used across the Synap docs, in one place.
Agent
An AI program that works on your pod through MCP or the Hub REST API, for example Claude Code, Codex or Cursor. Each connected agent has its own identity and key on the pod.
API key
A bearer token that lets an app, script or agent call your pod. It has scopes (which surfaces it may call) and a grant (what it may touch). See API keys.
Automation
A trigger (an event, a schedule (cron), a webhook, or a manual start) and the flow of steps it runs. Starting a playbook is one kind of step. See Automations and playbooks.
Channel
A conversation room. A session has a channel where you and your agents work on the same goal.
Control Plane
The Synap Cloud service at api.synap.live that creates and manages hosted pods, handles Synap accounts and
billing, and serves the marketplace catalog. A self-hosted pod can work without it.
Entity
One record in your pod: a person, a task, a note. Every entity has exactly one kind and zero or more roles.
Event
An append-only record of a change. The event log is the pod's audit trail.
Facet
Another name for a role attached to an entity.
Hub REST API
The pod's HTTP API under /api/hub, used by agents, scripts and integrations. Every pod publishes its OpenAPI
description at /api/hub/openapi.json. See Hub REST API.
Kind
What an entity is: person, task, event, note, article, decision, or a kind you define.
MCP
Model Context Protocol, the standard AI clients use to call tools. Every pod serves an MCP endpoint at /mcp.
See MCP.
Playbook
A session shape you can run again: a template for one kind of work, or a method a project runs as a track. Shown in the apps as a Template.
Pod
Your Synap server: the API, PostgreSQL database, search, file storage and sign-in, run as a set of containers. Self-hosted or on Synap Cloud.
Pod admin console
The pod's web console for its owner and admins, served on its own host name next to the pod (for example
pod-admin.example.com).
Project
Long-lived intent: something you are doing over weeks or months. Its progress is derived from the work it contains.
Proposal
A change an agent asks to make and that waits for a person. You review it and approve or reject it. Shown in the apps as a Request. Under the default write mode, an agent's creates and edits apply directly with Undo, and its deletions and structural changes become proposals. Deletes, admin actions and changes of scope are always proposals; an agent's own key can never approve one. See Validation policy.
Relay
The Synap mobile app (iOS and Android), where you approve proposals and capture things. It is in beta testing and not yet in the app stores. See Relay.
Role
A context an entity plays, such as client or blocker. A role is attached to the entity; it is not a second
entity. Also called a facet.
Session
A bounded unit of work toward one goal, with a scope you accept before it runs. It can stand alone or be filed at a track and stage. Shown in the apps as a Job.
Skill
A folder of instructions (a SKILL.md and reference files) that teaches an agent how to use Synap. See
Agent skills.
Stage
One step of a track. A session can be filed at a track and one of its stages. Shown in the apps as a Step.
Synap Cloud
Synap's hosted service: Synap runs your pod for you. See Self-hosted pod or Synap Cloud.
Track
A reusable method running inside one project, such as "Content" or "Build". It keeps the definition it started from. Shown in the apps as a Goal.
View
A projection over your records (a table, board, calendar, graph and so on). It never copies data.
Workspace
A space inside your pod that groups records, views and members, such as a business workspace and a personal one. Shown in the apps as a Space.
Write mode
How an agent's changes land, chosen when you connect the agent: Normal (the default: creates and edits apply directly with Undo; deletions and structural changes ask you), Creates only (creates apply with Undo; everything else asks you) or Ask first (every change asks you). See Validation policy.