Artifacts — Deep Notebook Docs

What Are Artifacts

Artifacts are structured outputs that blocks can generate beyond plain text. When you ask Deep Notebook to create something tangible — a document, a piece of code, a spreadsheet, or an image — it produces an artifact.

Artifacts are interactive. You can edit them, run code, copy data, and iterate on the output directly within the notebook.

Artifact Types

Deep Notebook supports four types of artifacts:

📄 Text Documents

Rich text documents with formatting. Perfect for reports, summaries, drafts, and any written content. Supports markdown-style formatting with headers, lists, and emphasis.

💻 Code

Syntax-highlighted code with execution support for Python. Generate scripts, analyze data with code, create visualizations with matplotlib. Run code directly in the notebook.

📊 Spreadsheets

Interactive spreadsheet editor for tabular data. Edit cells, copy as CSV, and work with structured data. Great for data transformation and analysis outputs.

🖼️ Images

Generated images from text descriptions. Create diagrams, illustrations, or visual content. View, copy, and iterate on generated images.

Creating Artifacts

Artifacts are created automatically when you ask for them in your instructions. Be explicit about what you want:

For documents: - "Write a summary of these findings as a document" - "Create a draft email responding to this thread" - "Generate a report with executive summary and key metrics"

For code: - "Write a Python script that analyzes this data" - "Create code to visualize these trends with matplotlib" - "Generate a script to process these CSV files"

For spreadsheets: - "Output this data as a spreadsheet" - "Create a table comparing these options" - "Format these results in a CSV structure"

For images: - "Create an image illustrating this concept" - "Generate a diagram showing this workflow"

Working with Artifacts

💻 data_analysis.py

◀ ▶ 📋 ▶ Run

import pandas as pd
import matplotlib.pyplot as plt
# Load and analyze the data
df = pd.read_csv('sales_data.csv')
monthly = df.groupby('month').sum()

Artifact actions:

•Version history — Navigate between previous and current versions using the arrow buttons
•Copy — Copy content to clipboard (text, code, or CSV for spreadsheets)
•Run — Execute Python code directly in the browser
•Edit — Modify the artifact content directly

Versioning

Every time you iterate on an artifact ("make this more concise" or "add error handling to this code"), Deep Notebook creates a new version. You can navigate between versions to compare changes or revert to a previous state.

Tips for Better Artifacts

Be specific about format. "Create a bulleted summary" vs "summarize this" produces better structure.

Provide context for code. Include what libraries are available, what the input data looks like, and what output format you need.

Iterate in conversation. After an artifact is created, ask for modifications: "Make the introduction shorter," "Add comments to the code," "Include a totals row in the spreadsheet."

Combine with integrations. Artifacts can be sent to connected apps: "Create this spreadsheet and save it to my Google Drive" or "Write this summary and post it to Slack."