Notebooks

Notebooks that show you how your analysis fits together.

See the whole picture

The graph canvas lays out every cell, every dependency, and every result—so you always know what happened, what connects to what, and what needs to rerun.

Q1 marketing review

Cells

Cell 3SQL8 rows · 41ms

-- Campaign data

SELECT channel, spend

FROM marketing.campaigns

Press ⌘ + ↵ to run⊙ Campaign data

Cell 10SQL9 rows · 53ms

-- CPA by channel

SELECT channel, SUM(spend)/COUNT(*) AS cpa

FROM marketing.campaigns GROUP BY channel

CHANNEL	CPA
Google	$22.58
Email	$3.85
Meta	$19.73

Press ⌘ + ↵ to run⊙ CPA by channel

Cell 11CHART

Cell 10

Google

Organic

Meta

Spend by channel

Cell 12CHART

Cell 10

Conversions trend

Insight EngineCampaign analysis

Break down CPA by channel

Google Ads has the lowest CPA at $22.58. I've created Cell 10 with the breakdown and added two charts derived from it.

Ask a question about your data…

DB: marketing-prod

Cell 10SQL5 rows53ms

1-- Campaign performance

2SELECT

3 DATE(created_at) as date,

4 COUNT(*) as conversions,

5 SUM(spend) as total_spend,

6 SUM(impressions) as total_impressions

7FROM public.marketing_campaigns

8GROUP BY DATE(created_at)

9ORDER BY date DESC

53ms

DATEdate	CONVERSIONS#	TOTAL_SPEND#	TOTAL_IMPRESSIONS#
2026-01-26	1	4200	18600
2026-01-25	3	3800	12400
2026-01-24	2	1200	9200
2026-01-23	2	2900	15800
2026-01-22	1	0	0

output: Cell 10rename(c_07d3ec4e63e8)⊙ used by 2+ attach cell

SQL that knows where it lives

Every SQL cell tracks its connection, execution time, row count, and downstream dependents. Change a query and see exactly what's affected.

Live results with timing and row counts.
Dependency badges show what uses this cell's output.
Run with ⌘+Enter—execution is instant feedback.

Charts that stay in sync

Chart cells derive from SQL results. When upstream data changes, charts update automatically—no copy-paste, no stale screenshots.

Bar, line, and more—configured visually, not in code.
Derived from specific cells with explicit lineage.
Always reflects the latest query results.

Derived: 1 inputCell 10

Cell 11CHART5 rows84ms

Type

Bar

X-Axis

Field

channel

Y-Axis 1

spendSum

Add Y-axis

Color / Group

Group by

None

spend

StackedLegend

Campaign Analysis6 cells

DAMKJL

DAMK

Cell 10SQL9 rows · 53ms

-- CPA by channel

SELECT channel,

SUM(spend) / COUNT(*) AS cpa

FROM marketing.campaigns

DA is editing

Cell 11CHART

DA — editing Cell 10MK — viewing Cell 11Agent — ready

Work together, in real time

Multiple people and the AI agent can work in the same notebook at the same time. See who's editing, follow their cursor, and let the agent build cells alongside your team.

Live cursors and presence indicators for every collaborator.
The AI agent shows up as a teammate—watch it create and run cells.
Follow mode to see exactly what a colleague is working on.

Analysis you can trust and rerun

Notebooks connect directly to your databases and semantic layers. Every result is reproducible, every workflow is auditable, and the AI agent works inside the same canvas as your team.