Connectors

A connector is a configured link between Trove and a data source. Each connector instance points to a specific source (e.g., “My Tech RSS” pointing at a particular feed URL, or “HN Top Stories” pulling your Hacker News saves).

Connector Types

Type	Description
hacker-news	Hacker News saved stories and top posts
rss	RSS and Atom feeds
arxiv	arXiv paper feeds
manual	Content saved via trove_save or the API
Custom	Any type you define via the API

Execution Modes

Cloud (CLOUD)

Trove runs the connector on a schedule. Configure a schedule like "every 6 hours" and Trove fetches new content automatically. No code required on your end.

Local (LOCAL)

Your app pushes data to Trove via the Sync API. Use this for data sources that require local access (e.g., a macOS app indexing local files) or custom logic you want to control.

Lifecycle

Status	Meaning
SETUP	Created but not yet fully configured
ACTIVE	Running normally. Cloud connectors sync on schedule; local connectors accept pushes.
PAUSED	Manually paused. No syncs occur until resumed.
ERROR	Last sync failed. Check errorMessage for details. The connector retries on the next scheduled run.

Schedules

Cloud connectors accept these schedule formats:

• "every 30 minutes"
• "every hour"
• "every 6 hours"
• "every 12 hours"
• "daily"

The cron system checks for due connectors every 5 minutes and enqueues sync jobs via a queue.

Sync Runs

Each sync creates a SyncRun record tracking:

• Status. RUNNING, SUCCESS, ERROR, or CANCELLED
• Documents synced. How many new documents were indexed
• Duration. How long the sync took
• Cursor. Position marker for incremental sync, so the next run picks up where this one left off

View sync history via the syncRuns field on any connector.

Manual Saves Connector

The first time you use trove_save (MCP) or the saveDocument mutation (GraphQL), Trove automatically creates a “Manual Saves” connector with type manual and status ACTIVE. All manually saved content goes here.

Connector Config

Each connector has a JSON config object with connector-specific settings (feed URLs, API credentials, filter parameters). Sensitive fields (API keys, tokens) are automatically redacted in API responses.

Managing Connectors

Mutation	Purpose
createConnector	Create a new connector with type, name, execution mode, schedule, and config
updateConnector	Change name, schedule, or config
pauseConnector	Pause syncs (status changes to PAUSED)
resumeConnector	Resume a paused connector (status changes to ACTIVE)
deleteConnector	Remove connector and all its documents (cascading delete)
syncConnector	Trigger an immediate sync (cloud connectors only)