Deepnote research: our notes on building agents
Get started

Online Python notebook built for collaboration

Run Python code online, explore your data, and share the results in one workspace. Write Python, query with SQL, and keep your code, charts and explanations together.

Start with a new project or bring an existing notebook. Your code runs in a hosted environment, with no local Python installation required. Invite your team to review the work and build on what you discover.

Online Python editor built for working with data

Write code, inspect a table, and see a chart beside the calculation that produced it. Deepnote’s online Python editor brings the steps of an analysis into one document, making it easier to explore a question and explain the answer.

Combine 25+ types of blocks

Every block type works with the same variables. Query connected data with SQL, transform the results with Python, and add an interactive chart without writing any code. Bring in input controls and agent blocks wherever your workflow needs them.

Run Python in the cloud

Choose a managed Python environment, including an option with common data science libraries preinstalled. Upload a dataset, configure a supported data connection, and add the packages your project needs.

Get help writing and understanding code

Ask Deepnote Agent to draft, edit or explain code, and suggest fixes when something goes wrong. Review its suggestions against your data and assumptions before running them.

G2 Leader in Python IDEs

Deepnote is recognized as a Leader and an Enterprise High Performer in G2’s Python Integrated Development Environments (IDE) category.

G2 Leader badgeG2 Enterprise High Performer badge
01/09/2026

Effortless Python App Deployment with Deepnote’s Intuitive Features

“When I was looking to deploy Python apps and dashboards for a course, I found that there were few solutions easy for students to deploy and share. Deepnote’s apps and dataframe visualizations are really neat and easy to deploy. I also really like the GitHub integration that makes it easy to share a project, rather than a single notebook.”

Read the full Deepnote review
Verified UserHigher Education
08/11/2025

Love that I can rapidly pull data from different sources and build visualizations.

“The workflow of integrating SQL and python is what I need for my data queries. I’m an embedded programmer and need to monitor the performance of various features across the devices in the field and this often involves pulling data from various data sources and then combining and correlating them. Doing this in a notebook that integrates with our data sources...”

Read the full Deepnote review
Verified UserLogistics and Supply Chain
06/24/2025

Powerful and versatile, good for both solo projects and client-facing work

“I’ve used Deepnote for everything from one-off dashboards for clients to big solo projects that involve stitching together data from multiple sources. It allows easy switching between Python and SQL in the same notebook (and maybe R? I haven’t tried that yet though!). It is also straightforward to integrate external data from Sheets or Snowflake...”

Read the full Deepnote review
Matthew T.Data Designer - Publishing

See a Python notebook turn data into an answer

Go from raw data to a shared insight in one workspace. Connect your data, analyze it with Python and AI, and turn the results into an analysis your team can use.

Weatherweather-forecast
weather_dbVariableobservations
1SELECT date, station,
2 temp_max, precip
3FROM daily_observations
4WHERE station = 'NYC'
5 AND date >= '2016-01-01'
6ORDER BY date
datedatetime64
stationobject
temp_maxfloat64
precipfloat64
02016-01-01NYC36.30.0
12016-01-02NYC34.00.3
22016-01-03NYC30.61.2

Built with Python in Deepnote

Explore real Python notebooks for sports analytics, music data analysis and drug discovery. Follow the code to see how each author approaches the work.

Code Python online with your team

Give teammates a shared place to understand the work, ask questions and build on the results.

Keep the reasoning beside the code

Review an analysis in the same project and leave comments. Record metric definitions, exclusions and assumptions so the next person can understand how you reached an answer.

Turn results into a useful output

Share the project when someone needs the underlying analysis. Publish a data app when they need a focused view of the results, with charts and input controls where useful.

Choose how others access the work

Set project access for the intended audience. Review the permissions available on your plan and give collaborators the access they need.

Turn your Python notebooks into an agent workspace

Deepnote Agent Workspace combines notebooks, data connections,
and business knowledge so agents can work with your data’s context.

Work from anywhere

Connect through Cursor, Claude Code, Codex, and other MCP-compatible tools. Access Deepnote notebooks and data within existing permissions.

Deploy always-on analytical agents

Schedule notebook workflows or trigger them through the API. Automate reports, data checks, and recurring analyses.

Build your context layer

Keep metric definitions, assumptions, analysis steps, and business context alongside your code. Turn trusted work into reusable skills, agents, and data apps.

Go deeper with Python in Deepnote

Find the documentation for your next step:

Python notebook FAQ







Start your next Python notebook with your team

Bring your code, explore data, and keep the result and the explanation in one place. Build an analysis that your team and agents can reuse.