Integrate external data sources into AI context with connection and sync status
| Prop | Type | Default | Description |
|---|---|---|---|
sources | ConnectorSource[] | required | Array of source entries with id, name, optional type, status ("connected" | "syncing" | "error" | "disconnected"), description, and lastSyncedAt |
onConnect | (sourceId: string) => void | required | Callback fired when the user connects a source |
onDisconnect | (sourceId: string) => void | required | Callback fired when the user disconnects a source |
onSync | (sourceId: string) => void | undefined | Optional callback fired when the user manually triggers a sync |
title | string | "Connectors" | Optional heading shown above the connectors list |
variant | "list" | "cards" | "list" | Visual presentation style for source items |
Connectors link external data sources to an AI workflow so generation can use real project context instead of only free-form prompts. They surface what sources are currently attached, whether each source is healthy, and when content was last synchronized. This pattern is especially useful in retrieval-augmented systems where the quality of output depends on the freshness and coverage of connected data.
Variants
- Source list — A compact list view showing each source, its connection state, and quick actions like connect, disconnect, and sync.
- Card catalog — Larger cards with provider metadata and richer status details for teams managing many integrations.
- Inline picker — A lightweight selector near the input that lets users toggle which connected sources should be used for the current run.
Use Cases
- Connecting docs and wiki platforms so AI can answer with product context
- Attaching storage providers for file-aware summarization and Q&A
- Linking databases for analytics-style prompts and report generation
- Monitoring sync health for enterprise knowledge integrations
- Scoping AI responses to selected trusted sources
Best Practices
- Show source status clearly (connected, syncing, error, disconnected) so users can trust retrieval context.
- Expose last sync timestamps and refresh controls to prevent stale context issues.
- Let users connect or disconnect sources without leaving the active workflow.
- Provide actionable error states with guidance when authentication or indexing fails.
- Support source-level permissions and visibility rules in team environments.