Deepnote Copilot
Deepnote AI Copilot is an AI-powered code suggestion tool developed by Deepnote that assists data scientists and analysts who work in Pytho...
Last verified:
What is Deepnote Copilot?
Deepnote Copilot (also called AI Copilot or Deepnote AI) is a contextual AI assistant designed for data scientists and analysts who work primarily in Python and other coding languages within notebooks. It provides lightning-fast, context-aware code suggestions as you type in code or SQL blocks, understanding the full scope of your notebook including previous code blocks, text blocks, files, and data. The tool helps users cut back on repetitive and boilerplate code, allowing them to focus on the big picture while AI handles tedious coding tasks.
Key features include AI-based code completion tailored to your notebook context, code generation from plain language instructions, code explanation for complex pieces of code, code editing and refactoring through natural language requests, and code debugging with immediate fixes. The tool also offers Auto AI mode—a fully autonomous agent that creates and executes text, code, and SQL blocks while self-correcting errors—and Generate mode that adds blocks without automatic execution. Deepnote AI understands projects, data warehouses, and metadata to provide precise, auditable assistance with transparent and reproducible results.
Deepnote Copilot is specifically built for power users: seasoned data practitioners including data scientists and analysts who do most of their work in coding languages. It's also suitable for coding beginners since the AI assists every step of the way. The tool partners with Codeium for blazing speed and great model performance with generous context windows. During their Beta programme, customers reported massive productivity boosts on par with Github Copilot experience in classic IDEs, now with the added benefits of collaborative notebooks.
Deepnote Copilot pricing
Pricing model: Freemium
Deepnote offers a freemium plan with three tiers: Free, Team, and Enterprise. The Free tier includes up to 3 editors, up to 5 projects, limited Deepnote AI, unlimited Basic machines with 5 GB RAM and 2 vCPU, and 7-day revision history. Free tier includes only 10 completions per month and 5 calls per month for AI features. The Team tier includes unlimited editors, unlimited projects, and enhanced AI capabilities with 30-day revision history. The Enterprise tier includes unlimited editors, unlimited projects, Plus machines (16 GB RAM, 4 vCPU), 24-hour SSH tunnel timeout, and unlimited revision history. You can start for free and upgrade as needs grow.
Deepnote Copilot pros
- Lightning-fast context-aware code suggestions
- Understands full notebook scope including previous blocks
- Reduces repetitive and boilerplate code significantly
- Works in both Python code blocks and SQL blocks
- Generate code from plain language instructions
- Provide concise explanations for complex code
- Edit and refactor code through natural language
- Debug code with immediate click-to-fix solutions
- Auto AI mode for fully autonomous task execution
- Auto AI self-corrects when encountering errors
- Generate mode retains control over block execution
- Creates multiple blocks at once from single prompt
- Deep understanding of data warehouses and metadata
- Precise and auditable AI assistance
- Transparent and reproducible AI results
- Partners with Codeium for blazing speed
- Generous context windows for relevant suggestions
- On par with Github Copilot productivity boosts
- Opt-in AI capabilities for user control
- Works remotely even when you close Deepnote
Deepnote Copilot cons
- Cannot yet use SQL with OAuth integrations
- Cannot create chart blocks yet
- Limited Deepnote AI on free tier
- Only 10 completions per month on free plan
- Auto AI requires workspace setting enablement
- Code completion needs individual workspace enable
- SQL block generation can be turned off in settings
- Collaborators can't stop your AI generation
Frequently asked questions about Deepnote Copilot
What is Deepnote Copilot?
Deepnote Copilot (also called AI Copilot or Deepnote AI) is a contextual AI assistant that provides lightning-fast, context-aware code suggestions for data scientists and analysts working in Python and other coding languages. It understands the full scope of your notebook including previous code blocks, text blocks, files, and data to deliver relevant coding advice.
How do I enable Deepnote AI in my workspace?
To enable Deepnote AI, navigate to Settings & Members in your workspace and click on AI. When AI code completion is turned on, every editor and admin user in the workspace will gain access to features like code generation, editing, explanation, and completions displayed in all projects.
What are the two modes of operation for Deepnote AI?
Deepnote AI operates in two modes: Auto AI is a fully autonomous agent that adds blocks, executes them, and self-corrects errors until achieving the desired outcome. Generate mode adds code and text blocks based on your instructions but waits for you to execute them, retaining control over block execution.
Can Deepnote Copilot work with SQL blocks?
Yes, Deepnote Copilot works in SQL blocks as well. SQL query completions utilize your notebook as context along with relevant SQL blocks using the same integration in other notebooks within your workspace, making suggestions more relevant as you do more work with the data source.
What is Auto AI and how does it differ from regular AI?
Auto AI is a fully autonomous mode that devises a plan for achieving your goals, then writes text, creates and executes code and SQL blocks until the task is completed. It self-corrects errors and can operate independently even if you close Deepnote. Regular Generate mode requires you to execute blocks manually.
How do I accept code suggestions from Deepnote Copilot?
Once enabled, code suggestions appear real-time as you type in code or SQL blocks. You can accept suggestions by pressing the Tab key. You can also prompt code suggestions by providing a textual description of what you want to achieve in a text block or comment.
Can I disable code completion for myself only?
Yes, you can toggle code completion for individual users by disabling it in the block actions menu. This affects code completions in all your notebooks but other workspace members will remain unaffected.
What guardrails exist for SQL usage by Auto AI?
Auto AI has two guardrails against SQL risks: AI is programmed to never execute write queries, and Deepnote has a detection mechanism that stops write queries from getting executed even if the AI generates one. Auto AI can only generate and run SELECT queries to fetch data.
How can I provide feedback on Deepnote AI quality?
Give thumbs up/down when reviewing AI-generated blocks. If an output didn't work, press 👎 to share more context in a survey window. You can also share ideas and suggestions on the Product Portal, or book time with the product team to be involved in shaping Deepnote AI's future.
What happens if I'm not satisfied with AI-generated output?
You can re-try code generation using the same prompt by clicking Try again, which places you back in the prompt bar to modify instructions for improved results. You can also remove all generated blocks with the Discard button. Deepnote AI has notebook access and is conversational, so you can tell it to Continue or Do task X.