MotherDuck + Querri
Querri reads the tables and views in your MotherDuck databases and copies them into its Library, where people who don't write DuckDB SQL can ask questions and build dashboards.
Querri only reads from MotherDuck and never changes your records. SOC 2 Type II, HIPAA and ISO 27001:2022 certified, and your data is never used to train a model.
What you can answer
MotherDuck suits people who are comfortable writing SQL, which leaves everyone else waiting on an analyst. Querri puts those tables behind dashboards the whole team can open and filter.
DashboardsJoining warehouse tables with CRM records usually means loading the CRM into MotherDuck first. Querri connects both and answers across them, with no extra pipeline to maintain.
HubSpotAnswering a one-off question in MotherDuck means writing the DuckDB SQL yourself. In Querri you ask in plain language, see the steps it took and follow up with the next question.
AI ChatWhat comes over
Querri copies this data into your Library on a schedule, next to every other system you connect, so one question can draw on all of it.
After the first load, each scheduled sync brings in only rows that are new or changed since the last one.
Querri reads from MotherDuck and never creates, edits or deletes anything there. Every answer runs on Querri's own copy.
Querri's own sharing and row-level access decide who sees which rows, in chat, dashboards, Claude, ChatGPT and the API.
How to connect
Say "connect MotherDuck" or pick it from the connector catalog.
Create an access token in MotherDuck's Settings and paste it into the secure form; a Read Scaling token is enough. Name one database to open, or leave the field empty to see all of yours.
Pick the data you want and how often it refreshes, then approve the first load.
The data lands in your Library, ready for chat, dashboards and scheduled reports.
FAQ
Paste a MotherDuck access token into the secure form, which the Connector Agent never sees. A Read Scaling token is enough because Querri only reads. You can name one database to open; leave it empty and Querri lists all of your databases.
Tables and views, chosen by schema. Schemas appear as database.schema, for example my_db.main. MotherDuck's shared sample_data database and DuckDB's system catalogs are left out of the list, so you only see your own data.
On the schedule you set. Full refreshes reload whole tables daily, weekly or monthly. Incremental syncs run hourly, daily or weekly and fetch only rows beyond the last value of a cursor column, such as an event timestamp, which suits large append-only event tables.
No. This connector opens MotherDuck databases only, and Querri locks the DuckDB session so it can't read local files or install extensions. To analyze a local .duckdb file, load it into MotherDuck first, or export its tables to CSV and upload them to Querri.
No. Querri only reads from MotherDuck. It never creates, edits or deletes records there, and every answer runs on Querri's own copy of the data.
Yes. Querri is SOC 2 Type II, HIPAA and ISO 27001:2022 certified, credentials are stored encrypted, and your data is never used to train a model.
Keep exploring
Connect MotherDuck, add the rest of your systems, and get answers your whole team can check.
