What is a connector?

A connector reads from a source system and writes structured data into a Curiosity Workspace, keeping the two in sync over time.

Built-in integrations (no code) cover common sources: SharePoint, Google Drive, Confluence, SQL databases, file shares, and others — configurable in Settings → Integrations.

Write a custom connector when:

• Your source isn't in the built-in list
• You need control over how records map to the graph
• You need to mirror the source's access permissions
• You're combining data from multiple sources into one schema

---

Every connector does five things:

1. Authenticate — API token scoped to ingestion
2. Register schemas — declare node and edge types (once, at startup)
3. Read from source — initial full load, then incremental deltas
4. Upsert — TryAdd nodes, Link edges, set permissions
5. Commit — in bounded batches; record a cursor for the next run

The workspace then handles text extraction, NLP, and indexing automatically.

---

→ Connectors reference