Documentation — DataPyn

Introduction

DataPyn is an IDE for people who work with data every day.

Pynia is the AI built into DataPyn.

Mixed blocks

SQL and Python blocks in one file.

Per-block connections

Each SQL block can use a different connection.

Pynia

Chat with your data stack.

Monaco + offline SQL

Schema-aware editing, cross-database refs.

Installation

Windows (recommended)

Download DataPyn-Setup.exe from GitHub Releases.

DataPyn-Setup.exe · All releases

From source

For contributors or custom setups on Windows:

git clone https://github.com/natharuc/datapyn.git
cd datapyn
.\scripts\install.bat
.\scripts\run.bat

First connection

You need at least one saved connection.

Open connections panel.
New connection.
Test and save.
Connect to session.

Schema loads after connect.

Code blocks

A session is a stack of blocks.

Add or switch blocks

Toolbar buttons
Shortcuts
Language picker

Named results

Name SQL blocks for Python.

-- block: sales
SELECT region, SUM(amount) AS total FROM orders GROUP BY region

# Python
print(sales.sort_values("total", ascending=False).head())

SQL editor

Monaco editor with SQL intelligence.

Offline intelligence

Schema-based autocomplete without internet.

Cross-database references

other_db..table notation.

SELECT p.nome, v.valor
FROM minha_base..vendas v
JOIN outro_catalogo..pessoa p ON p.id = v.pessoa_id

Pynia ghost text

Inline AI suggestions while typing.

Databases

Supported connectors:

Database	Notes
SQL Server	Windows Auth, SQL Auth
MySQL	Multiple databases
PostgreSQL	Schemas
MariaDB	MySQL-compatible
SQLite	Local file
Databricks	SQL warehouse

Execution

Common shortcuts:

Shortcut	Action
F5	Run selection or all
Ctrl+F5	Run all in order
Shift+Enter	Run and advance

SQL → Python

SQL results in Python:

SELECT category, SUM(revenue) AS total FROM sales GROUP BY category

# df or named block
df.plot(kind="bar", x="category", y="total", legend=False)

Visualization

matplotlib, seaborn, plotly…

Pynia datapyn_chart tool.

Import & export

Drag & drop

CSV, Excel, JSON.

Export

Results grid
Python script
Workspace .dpw

What is Pynia

Pynia is the chat layer in DataPyn.

Context-aware help for SQL, Python, schema.

Billing is on you

DataPyn is free; providers may charge.

OpenAI

Claude (Anthropic)

Open Router

GitHub Copilot

Connectors & settings

Settings → Pynia

GitHub Copilot
OpenAI
Claude
Open Router

Separate chat and autocomplete models.

Chat

Open Pynia panel.

#block and #tab references.

Switch connector and model.

Inline autocomplete

Ghost text in blocks.

Same connector token as chat.

Tools Pynia can call

Just ask naturally.

Tool	What it does
`datapyn_snapshot`	Workspace snapshot
`datapyn_inspect`	Inspect block/variable
`datapyn_query`	Quiet SQL/Python run
`datapyn_run`	Visible run/write
`datapyn_edit`	Edit blocks
`datapyn_blocks`	Create/focus tabs
`datapyn_database`	Connections & schema
`datapyn_chart`	Charts
`datapyn_subagent`	Parallel explore
`datapyn_notify`	Toasts

Example prompts

Day-one ideas:

Explore data

What tables are in this database?
Summarize revenue by region for last quarter
Find duplicates in the orders table

Charts

Bar chart of the last query by category
Plot monthly trend from block sales

Fix & automate

Why did this SQL block fail?
Refactor this query to use a CTE
Add a Python block that exports top 100 rows to Excel

Keyboard shortcuts

Execution

F5	Run
Ctrl+F5	Run all
Shift+Enter	Run + next
Ctrl+Enter	Run all blocks

Editing

Ctrl+D	Duplicate line
Ctrl+/	Toggle comment
Ctrl+F	Find
Ctrl+H	Replace
Ctrl+G	Go to line

Tabs

Ctrl+T	New tab
Ctrl+W	Close tab
Ctrl+Tab	Next tab

Blocks

Ctrl+Shift+S	New SQL block
Ctrl+Shift+P	New Python block
Ctrl+Shift+B	Add block
Alt+↑/↓	Move block

Pynia & connections

Ctrl+Shift+C	Connections panel
Toolbar / menu	Open Pynia

FAQ

Windows Auth?

Choose Windows authentication.

Multiple DBs?

Per-block connection selector.

Password storage?

AES-256 local encryption.

Open Pynia?

Toolbar, menu, or View dock.

Which provider?

Copilot if you have it; else API keys.

Export .py?

Menu → Export → Python script.