Documentation
Getting started
Your context in every connected AI app.
Connect your appClient-specific setup with steps and commands.
Accounts, apps and agentsAgents under your account with selected context, their own tasks and controlled integrations.
Connect headless agentsPersonal MCP access with a device code, without an SSH tunnel.
Organize contextKnow when to use skills, facts, wikis and workspaces.
Sharing & accessCollaborate with explicit access grants.
Action itemsHow invitations, waiting records and messages reach your connected AI apps.
Read-only contextRead sources without giving the agent write authority.
Create your own pluginsDesign tables and forms visually and work directly with saved contacts.
Use the Plugin HubFind, install, share, rate and update plugins privately, within an organization or publicly.
Plugins
Capture commitments, review your inbox and track projects.
CalendarPlan appointments and work blocks, optionally linked to tasks.
MessagesChannels, approvals, replies and optional task context.
SecretsStore, rotate and export encrypted credentials.
IntegrationsConnect provider accounts and approve external tools.
Context explorerRead context files, review drafts and download files.
MeetingsKeep notes, transcripts and reviewed outcomes together.
Conversations and sessionsImport transcripts, continue across clients and share through buckets.
Plugin SDK
From Hello World to a reviewed package: schemas, runtime, MCP and web views.
API referencePublic SDK exports, host interfaces, codecs, permissions and query contracts.
Dependencies & contractsIndependent plugins, optional capabilities and versioned public schemas.
Reviewed commands & audiencesExact command approval, fixed audiences and isolated client credentials.
Agent accessExperimental contract: trusted actors, scoped credentials and authenticated work sessions.
Credentials & external effectsEncrypted payloads, credential bindings, OAuth handoffs and reviewed provider calls.
Test & releaseContract tests, authorization, MCP, browser workflows and external model evaluations.