Integrations overview
Pull live data from external services into your artifacts, or push artifacts out to where your team already works.
Available
Section titled “Available”| Provider | What it gives you |
|---|---|
| Google Sheets | Read spreadsheet data with OAuth |
| Google Analytics | GA4 reports via pasted service-account key |
| Google Ads | Campaign spend, clicks, conversions |
| Facebook Ads | Meta campaign insights |
| Amazon S3 | List and read CSV, JSON and text files from a bucket |
| Supabase | Read-only SQL on your Supabase database |
| Amazon Redshift | SQL on a Serverless workgroup or cluster |
| Amazon Aurora | SQL on Aurora PostgreSQL or MySQL |
| Shopify | Products, orders, inventory |
| Slack | Deliver artifacts to a channel or DM |
| Tienda Nube | LATAM e-commerce |
| GitHub | Backup and export |
| CORS proxy | Call allow-listed external APIs from the page |
Two kinds of connection
Section titled “Two kinds of connection”- Per-artifact / OAuth-in-page — e.g. Google Sheets: the visitor (or owner) authorizes, and the artifact fetches data. Simple, no admin setup.
- Workspace-level — platform connectors and REST/warehouse connectors are
defined once per workspace. Artifacts reference them by name via
sdk.connection('name'). Most platform connectors now use a bring-your-own credentials model: paste your provider token or service-account key in the workspace admin catalog, Test before saving, then copy thesdk.connection('name')snippet.
The connector catalog is always visible in workspace admin — browse providers, connect, and test without hunting through settings.
Token safety
Section titled “Token safety”Credentials are never exposed to the page — calls are proxied through the Worker,
and secrets are encrypted at rest. Use POST /v1/workspaces/{id}/connections/test
to verify pasted credentials before saving. For LLM API keys (OpenAI, Anthropic,
…), use the secrets proxy in the AI agent guide.