Weekly Airbyte data catalog in Notion for business teams

Every Monday at 6am, refresh a Notion "Data Catalog" database so non-data teammates in marketing, sales, and finance have a self-serve view of everything flowing through our Airbyte pipelines, without having to ping the data team.

Start by pulling the current picture from Airbyte. Use Airbyte List Connections to enumerate every pipeline, Airbyte List Sources and Airbyte List Destinations to resolve each connection's source system and destination system by id, and Airbyte Get Connection Details on each connection to read its schedule and the list of streams that are actually enabled. For the most recent successful sync per connection, call Airbyte List Jobs filtered by that connection id with jobType=sync, then pick the newest entry whose status is succeeded.

Then, for each Airbyte connection, use Notion Create a Page inside the pre-existing Data Catalog database. The database already exists and its id will be provided at setup — do not create a new database, only add pages to that one. Populate the following properties on each page:

Title: a short human-readable name like "HubSpot → Snowflake (Marketing)". Source system: the source connector name (e.g. HubSpot, Salesforce, Stripe). Destination system: the destination connector name (e.g. Snowflake, BigQuery, Redshift). Sync schedule: rendered in plain English like "Every 6 hours", "Daily at 2am UTC", or "Manual" — not a cron expression. Last successful sync time: the timestamp of the most recent succeeded sync job; if the connection has never succeeded, leave it empty and note "never synced". Streams synced: the enabled stream names as a comma-separated list. Summary: a two-to-three sentence plain-English description of what data this pipeline moves and who typically consumes it — for example, "Pulls contacts and deals from HubSpot into Snowflake every six hours. Feeds the marketing attribution dashboard and the weekly AE pipeline review."

When writing summaries, remember the audience is non-technical. Avoid data-engineering jargon like CDC, incremental, namespace, or upsert. Frame each summary as (1) what data flows from where to where, (2) how often, and (3) who or what downstream tool tends to use it. If you don't know the consumer, make a reasonable guess based on the source system (e.g. HubSpot data is usually consumed by marketing and sales; Stripe data by finance; product event data by product and analytics).

If a Notion page already exists in the catalog database for a given connection (match on the connection id or title), overwrite it with the fresh values so the catalog stays current instead of accumulating duplicates. If a connection has been deleted from Airbyte since the last run, leave any existing Notion page alone but do not create a new one for it.