MotherDuck

Revenue, experiments and deals in MotherDuck, answered for the whole team.

Dot writes DuckDB SQL, runs it read-only and shows the query.

Threadanalytics

Priya Shah9:41 AM

@Dot How much of our MRR growth this year came from upgrades?

1 reply

DotAPP9:42 AM

10 steps · 53s · Access Online

Upgrades added $29K, 43% of the growth.

Source: MotherDuck’s billing.mrr.

Chart

Threadanalytics

Daniel Weber9:41 AM

@Dot Did our new pricing page lift trial starts? Compare both versions since launch.

1 reply

DotAPP9:42 AM

7 steps · 36s · Access Online

Yes: 18% more trial starts with version B, at 99% confidence.

Source: MotherDuck’s web.experiments.

Here is the data.

VersionVisitorsTrial starts
A24,1803.9%
B24,0204.6%

Threadanalytics

Lisa Park9:41 AM

@Dot Which of our deals over $20K have had no activity in two weeks?

1 reply

DotAPP9:42 AM

9 steps · 53s · Access Online

11 deals worth $480K. Four belong to the same rep.

Source: MotherDuck’s crm.deals.

Here is the data.

DealValueQuiet for
Arden renewal$86K19 days
Halcyon$64K16 days
Moreau$41K15 days

How GetDot works with MotherDuck

How Dot connects and what it reads

A MotherDuck connection string looks like md:my_db?motherduck_token=…, and one token reaches every database on its account, so name the databases and schemas Dot may use.

Dot reads tables and views from DuckDB’s catalog and writes DuckDB SQL, and read queries open the connection read-only.

Permissions and lockdown

Dot sees what its token sees, narrowed by your database and schema lists. For per-person rows, Dot’s row-level rules (Team plan) filter each query by the asker’s groups or attributes.

On every MotherDuck connection Dot turns off community extensions and DuckDB’s local file system, locks that configuration, and checks each query before it runs, so no question can reach files on the machine Dot runs on.

Dot and MotherDuck’s own AI

As of October 2026, MotherDuck offers a remote MCP server (fully managed, with read and read-write query tools) and a local one (self-hosted, read-only by default), plus AI functions in SQL: prompt() sends text to an LLM row by row, and prompt_sql turns a question into SQL.

Use those when one person wants their own assistant on MotherDuck, or LLM calls inside SQL, like tagging reviews. Use Dot when a whole team asks in Slack, Teams or email and needs shared definitions, scheduled reports and per-person rows across MotherDuck and your other sources.

Questions

Connect your data.

Start for Free