Jupyter AI is a powerful generative AI extension integrated into the Syntasa Notebook environment. It brings Large Language Models (LLMs) directly into your data science and engineering workflows, enabling users to generate code, debug errors, explain logic, and summarize data—all without leaving the notebook interface.
By combining conversational AI with cell-level execution, Jupyter AI enhances productivity while preserving reproducibility and governance within Syntasa.
Core Interfaces
Chat Panel
The Chat Panel is available in the left-hand sidebar of the Syntasa Notebook UI. It provides a persistent, thread-based conversational interface with the AI and maintains awareness of your current notebook context.
Key Capabilities
Code Generation & Refactoring
Request new functions, optimize existing logic, or translate code between languages (for example, SQL to PySpark).Contextual Commands
Jupyter AI supports slash commands for common workflows:/generate [prompt]– Create a new notebook from a natural language description/learn– Index files in the current directory to enable context-aware answers/ask [question]– Query the indexed files to answer project-specific questions/fix– Analyze the most recent error or traceback and suggest a fix
Selection Awareness
Highlighting code in a notebook cell allows the Chat Panel to explain, modify, or refactor only the selected portion.
Configuration and Setup
Provider Selection
Before using Jupyter AI, a language model provider must be configured:
Open Jupyter AI Settings (gear icon) in the Chat Panel
Select a Language Model Provider, such as:
OpenAI
Anthropic
Amazon Bedrock
Azure OpenAI
Select an Embedding Model, which is required for the
/learnand/askfeatures
Authentication
Authentication is handled securely using one of the following approaches:
API Keys
Enter provider-specific API keys directly in the Jupyter AI settings panel.Syntasa Secrets (Recommended for Enterprise)
API keys can be preconfigured as environment variables (for example,OPENAI_API_KEY) within Syntasa, enabling seamless access without manual user configuration.
Security and Data Privacy
Syntasa’s Jupyter AI integration is designed with strong data governance controls:
Explicit Data Transmission
Only the data explicitly included in a chat prompt or%%aicell is sent to the AI provider.Local Indexing
The/learncommand creates a local vector index within the Syntasa environment. This index is not shared externally. Only relevant snippets retrieved during a query are sent to the LLM.Compliance Considerations
Users should ensure that interactions with third-party LLMs align with organizational policies on data privacy, residency, and compliance.
Best Practices
Be Specific in Prompts
Clearly specify libraries, frameworks, and expected output formats (for example, “Use pandas” or “Return PySpark code”).Review Before Execution
Always review and validate AI-generated code before running it on production datasets.Manage Token Usage
Avoid passing extremely large datasets or long files directly into prompts, as this may exceed model context limits.
Summary
Jupyter AI in Syntasa notebooks enables a seamless fusion of generative AI and data engineering workflows. By combining conversational assistance, cell-level AI execution, and enterprise-grade security controls, it empowers teams to move faster while maintaining reliability, transparency, and compliance.