SaaS
Nothing to install. Start querying today.
Thunderduck SaaS runs on AWS and we operate all of it: the engine, the catalogs, upgrades and scaling. Sign up, connect a source and run your first query today. Every major feature ships here first.
Fully managed
We run the platform. You run the questions.
Compute on demand
Each query gets its own worker, sized from its plan. Nothing sits idle and there is no cluster to size.
Upgrades handled
New releases roll out continuously. You never schedule an upgrade or patch a server.
Security built in
Single sign-on, roles per catalog, and credentials kept in a secrets manager, never in plain text.
Pay for what runs
Every query is priced before it starts. You pay for the seconds it used, not for a warehouse left on.
First to get everything
Every major feature lands in SaaS first.
New capabilities ship to SaaS as soon as they're ready, then follow to BYOC and on-prem. Recent releases that started here:
- The MCP server for Claude, ChatGPT and Cursor
- DBT pipelines from Git, with freshness and alerts
- LakeOps for Iceberg compaction and snapshot cleanup
- AI error triage and the semantic layer
Runs on AWS
Operated by the Thunderduck team
- Connects to your databases, warehouses and buckets where they are
- Reads Iceberg on Glue, Nessie, Cloudflare and more
- Moves to BYOC later without changing a query
