Data Bengal
Get early access
Built for freelancers, consultants, small businesses & students

Data Bengal: AI Integrated SQL, Python, Smart Sheets, Visuals

Use any tool on its own, or all five together. Data flows between them automatically.

  • One workspace, shared data across every tool
  • Built-in AI — 100M tokens/month included. No code required
  • Share a link. No login, no license for your collaborators
0 / 150 founders' spots claimed · 150 left
Explore the workspace

Five tools. One workspace.

Each tool works on its own or with the others. Data flows between them automatically.

A SQL editor that knows your schema.

A real SQL editor in your browser. Works with DuckDB, Postgres, MySQL, BigQuery, and Snowflake.

  • Schema-aware autocomplete for tables, columns, and joins
  • AI writes queries from plain English
  • Connect any database, S3 file, or URL
  • Fast queries on multi-GB data
  • Results flow into Python, the spreadsheet, and dashboards
acme.duckdb connected
-- AI: "top regions Q1 with YoY"
SELECT region, SUM(amount) AS revenue,
  COUNT(DISTINCT customer_id) AS customers
FROM orders WHERE ts >= '2026-01-01'
GROUP BY 1 ORDER BY 2 DESC;
✓ 5 rows · 384 ms · 2.1 GB scanned
regionrevenuecustomers
NA-East$4,128,40238
EMEA$3,891,01531
NA-West$3,204,77829
APAC$2,440,99122
LATAM$1,108,54012

Jupyter notebook. Your df, already loaded.

Jupyter cells in your browser. Your SQL result is already loaded as df. No install.

  • Real Jupyter cells with rich outputs
  • SQL result loaded as df — no exports
  • AI writes cells, explains errors, suggests fixes
  • Runs in your browser — no install
  • Send output to a chart or spreadsheet
churn-analysis.ipynb Python 3.11
In [1]:
# df from SQL block above — no import needed
df.shape, df.columns.tolist()
Out[1]:
((12402, 8), ['order_id', 'ts', 'region', 'tier', 'amount', 'channel', 'days_since', 'churned'])
"Fit a churn classifier on tier + days_since"
In [2]:
from sklearn.linear_model import LogisticRegression
X = df[['tier_e', 'days_since']]
model = LogisticRegression().fit(X, df['churned'])
df['churn_score'] = model.predict_proba(X)[:, 1]
Out[2]:
model accuracy: 0.834 · added column 'churn_score' to df

A spreadsheet, linked to your data.

A familiar spreadsheet, linked both ways to your data. Edit a cell, the query updates. Edit the query, the sheet refreshes.

  • Familiar formulas: SUM, AVG, IF, VLOOKUP, ranges
  • Two-way linked to SQL and Python
  • Use the sheet as a SQL table in one click
  • AI fills columns and writes formulas
  • Export to .xlsx, .csv, or .ipynb
budget-2026.xlsx live-linked
File Home Insert Formulas Data View
Paste ▾ Clipboard
B I U Font
$ % , Number
Σ Sum ▾ Editing
Ask AI
D2
ƒx
=B2/SUM(B$2:B$6)
A
B
C
D
1
region
Q1_revenue
customers
% of total
2
NA-East
$4,128,402
38
27.9%
3
EMEA
$3,891,015
31
26.3%
4
NA-West
$3,204,778
29
21.7%
5
APAC
$2,440,991
22
16.5%
6
LATAM
$1,108,540
12
7.5%
7
Total
$14,773,726
132
100%
Sheet1× Forecast +

Dashboards, built by AI. Published to your team.

Describe what you want — AI builds the charts, filters, and layout. Publish for your team or spin up an ad-hoc report in seconds.

  • "Build me a Q1 sales dashboard" — AI picks the charts, KPIs, and filters
  • Ad-hoc reports for one-off questions; full dashboards for the work you do every week
  • Organized by domain — Sales, Finance, Marketing, Ops, Product
  • Publish & navigate. Your team browses a workspace, not a folder of files
  • Cross-filter across every chart, drill from any mark into the underlying SQL
  • Same data as your SQL queries, Python notebooks, and spreadsheet
Acme · Workspace 8 published
Workspace/Sales/Q1 Revenue Review Published
Q1 Revenue Review
✦ "build me a Q1 sales dashboard with revenue by region, tier breakdown, and YoY" AI-built
⌖ Region: All ⌖ Tier: All ⌖ Q1 2026 ⌖ Channel: Direct
$14.7M
Revenue
+18%
YoY growth
132
Customers
Revenue by Region stacked by tier · click to drill
NA-East
$4.1M
EMEA
$3.9M
NA-West
$3.2M
APAC
$2.4M
LATAM
$1.1M
Daily revenue (Q1)
$163K▲ 12%
New customers (Q1)
26▲ 8%
Enterprise Mid-market SMB 12,402 rows · 384 ms · refreshed 2m ago

One link. No login. No license.

Pick what to share. Send a link. They open it in any browser.

  • Three share scopes: workbench, dashboard, or single viz
  • Private links with optional password and expiry
  • Custom subdomain on team plans
  • Standalone HTML export — one self-contained file
  • Download as .ipynb, .sql, .xlsx, or .csv
Publish & share private link
Whole workbench
Editable · all blocks
Dashboard only
Viewer · with filters
Just one viz
Embed-sized
https://databengal.app/d/acme-q1-9k2zT
Dashboard mode. Viewers see the dashboard with filters. No login. No license.
Or download as
⬇ notebook.ipynb ⬇ queries.sql ⬇ workbook.xlsx ⬇ result.csv