BigQuery
BigQuery is on the roadmap. Agents will explore datasets and table schemas, write SQL, dry-run it to show the bytes scanned, and return the answer as a summary or a chart — while anything that writes a table or changes a dataset needs explicit human approval.
What your agents can do with BigQuery
- Explore datasets and table schemas
- Write and run read-only SQL
- Dry-run queries to estimate cost
- Create or replace tables (with approval)
How to connect BigQuery to Velaris
- 1
Get early access
Join the waitlist — we open access in waves.
- 2
Request BigQuery
Vote for the BigQuery connector in the community.
- 3
Get notified
We'll ping you the moment BigQuery goes live.
Popular ways to use it
BigQuery + Velaris — common questions
How do I stop an agent running a huge, expensive query?
Every agent query is dry-run first so you see the bytes it would scan, and you can set maximum bytes billed on the connection — over the cap, the query is refused.
Can an agent create or overwrite a BigQuery table?
Only with approval. Reads run on their own, but CREATE, INSERT and any destination-table write is staged with the exact SQL and waits for your sign-off.
Which IAM role should the connection use?
BigQuery Data Viewer plus Job User covers every automatic read. Add write roles only if you want the approval-gated table actions to be available.