How to Get Started with Quadratic AI Spreadsheet

Spreadsheets handle most business data work, but they break down fast: formulas become unreadable, refreshing a data source requires manual export and paste, and anything beyond a pivot table requires a separate tool. Quadratic is a browser-based spreadsheet that lets you run Python and SQL directly inside cells, connect live to a database, and use AI to write and debug code for you. You do not need to know how to code to get useful output.

What Quadratic Actually Is

Quadratic is built on an infinite canvas. Each cell can hold a standard formula, a Python snippet, a SQL query, or the result of an AI prompt. The output of any code cell renders directly into the spreadsheet grid, so charts, tables, and transformed data sit alongside the rest of your work in one file.

A team using Quadratic to track weekly sales could write one SQL block that pulls live data from a Postgres database, a Python block that calculates week-over-week change, and a formula column for formatting, all in a single spreadsheet that refreshes on demand. No export. No paste. No separate dashboard tool.

Creating Your First File

Go to quadratichq.com and sign up for a free account. The free plan includes full feature access with limited AI usage per month. Once logged in, click New spreadsheet to open the editor.

The canvas looks like a standard grid, but the toolbar includes a Code button and an AI button at the top. You can type into cells normally, or open a code panel for any cell by pressing = and then selecting Python, SQL, or JavaScript from the dropdown that appears.

Importing and Connecting Data

Quadratic supports three ways to get data into a file.

CSV upload is the fastest starting point. Drag a CSV file onto the canvas and Quadratic places it as a table. From there you can click any cell and open a Python block to filter, group, or transform that data using Pandas.

Database connections require you to go to Settings in the left sidebar and click Connections. Supported databases include PostgreSQL, MySQL, Snowflake, and BigQuery. Enter the hostname, port, database name, and credentials. Quadratic stores the connection at the account level, so you can reuse it across files.

Once a connection is saved, open any cell, select SQL as the cell type, and write a query. For example:

SELECT region, SUM(revenue) AS total
FROM sales
WHERE date >= '2026-01-01'
GROUP BY region
ORDER BY total DESC

The result renders as a table starting from the cell you placed the code block in. The query runs live each time you open the file or press the refresh button on the cell.

API fetch is available through Python. In a Python cell, use the requests library to pull JSON from any API endpoint, parse it, and return a dataframe. Quadratic has requests and pandas pre-installed.

Using AI to Write Code

The AI panel in Quadratic accepts plain English instructions. Click a cell, open the code panel, and instead of typing code directly, click Ask AI. Type what you want: "Group this table by country and sum the revenue column." Quadratic writes the Python or SQL, runs it, and shows the output.

You can edit the generated code directly or ask the AI to revise it. This loop works well for analysts who know what they want from the data but are not comfortable writing Pandas transformations from scratch.

One practical example: an ops manager tracking vendor invoices uploaded a CSV of 2,400 rows. They typed "Show me the top 10 vendors by total invoice amount, last 90 days only." Quadratic returned a ranked table in the grid with no formula writing required.

Visualizing Results

To create a chart, select the range you want to plot, click Chart in the toolbar, and choose a type. Bar, line, scatter, and pie are all available. The chart embeds in the canvas next to the data it references and updates when the source data changes.

For custom charts, a Python cell can use matplotlib or plotly. Quadratic renders the output figure directly in the cell's canvas area.

Sharing and Collaboration

Files are shared via link. Click Share at the top right and set the permission level: view only or can edit. Multiple users can work in the same file simultaneously, with changes syncing in real time, similar to Google Sheets.

Files are stored in Quadratic's cloud. There is no local export required to share work, though you can export the current state as CSV from the File menu.

Limitations to Know

Quadratic's free plan caps AI usage. Heavy users running AI-generated code blocks throughout a workday will hit the limit and need to upgrade. The Pro plan costs $29 per user per month as of early 2026.

Database connections require network access from Quadratic's servers to your database. For databases behind a firewall, you will need to allowlist Quadratic's IP ranges, which are listed in their documentation under Connections.

Python execution runs in a sandboxed WebAssembly environment. This means most standard libraries work, but anything requiring system-level access or unpublished packages will not run.

If you want to go further without managing code at all, tools like VSLZ let you upload a file and get analysis from a single prompt, with no code cells to manage.

Getting the Most from Quadratic

Start with a CSV you already work with regularly, something you currently analyze in a standard spreadsheet. Import it and try replacing two or three of your most complex formulas with Python or SQL equivalents using the AI panel. The goal is to verify that the output matches what you already know, so you build confidence in the tool before relying on it for decisions.

Once you have a working file, connect it to a live database source so it pulls fresh data on demand. That single change replaces an entire manual export-and-paste workflow.

Quadratic suits analysts who want more control than a BI tool offers and more speed than writing standalone scripts provides. It fills the gap between a spreadsheet and a notebook without requiring a choice between them.