Marimo Reactive Notebook Prompt Library

Starter codebase to use Marimo reactive notebooks to build a reusable, customizable, Prompt Library.

Take this codebase and use it as a starter codebase to build your own personal prompt library.

Marimo reactive notebooks & Prompt Library walkthrough

Run multiple prompts against multiple models (SLMs & LLMs) walkthrough

1. Understand Marimo Notebook

This is a simple demo of the Marimo Reactive Notebook

• Install hyper modern UV Python Package and Project
• Install dependencies uv sync
• Install marimo uv pip install marimo
• To Edit, Run uv run marimo edit marimo_is_awesome_demo.py
• To View, Run uv run marimo run marimo_is_awesome_demo.py
• Then use your favorite IDE & AI Coding Assistant to edit the marimo_is_awesome_demo.py directly or via the UI.

2. Ad-hoc Prompt Notebook

Quickly run and test prompts across models

• 🟡 Copy .env.sample to .env and set your keys (minimally set OPENAI_API_KEY)
  • Add other keys and update the notebook to add support for additional SOTA LLMs
• 🟡 Install Ollama (https://ollama.ai/) and pull the models you want to use
  • Update the notebook to use Ollama models you have installed
• To Edit, Run uv run marimo edit adhoc_prompting.py
• To View, Run uv run marimo run adhoc_prompting.py

3. ⭐️ Prompt Library Notebook

Build, Manage, Reuse, Version, and Iterate on your Prompt Library

• 🟡 Copy .env.sample to .env and set your keys (minimally set OPENAI_API_KEY)
  • Add other keys and update the notebook to add support for additional SOTA LLMs
• 🟡 Install Ollama (https://ollama.ai/) and pull the models you want to use
  • Update the notebook to use Ollama models you have installed
• To Edit, Run uv run marimo edit prompt_library.py
• To View, Run uv run marimo run prompt_library.py

4. Multi-LLM Prompt

Quickly test a single prompt across multiple language models

• 🟡 Ensure your .env file is set up with the necessary API keys for the models you want to use
• 🟡 Install Ollama (https://ollama.ai/) and pull the models you want to use
  • Update the notebook to use Ollama models you have installed
• To Edit, Run uv run marimo edit multi_llm_prompting.py
• To View, Run uv run marimo run multi_llm_prompting.py

5. Multi Language Model Ranker

Compare and rank multiple language models across various prompts

• 🟡 Ensure your .env file is set up with the necessary API keys for the models you want to compare
• 🟡 Install Ollama (https://ollama.ai/) and pull the models you want to use
  • Update the notebook to use Ollama models you have installed
• To Edit, Run uv run marimo edit multi_language_model_ranker.py
• To View, Run uv run marimo run multi_language_model_ranker.py

General Usage

See the Marimo Docs for general usage details

Personal Prompt Library Use-Cases

• Ad-hoc prompting
• Prompt reuse
• Prompt versioning
• Interactive prompts
• Prompt testing & Benchmarking
• LLM comparison
• Prompt templating
• Run a single prompt against multiple LLMs & SLMs
• Compare multi prompts against multiple LLMs & SLMs
• Anything you can imagine!

Advantages of Marimo

Key Advantages

Rapid Prototyping: Seamlessly transition between user and builder mode with cmd+. to toggle. Consumer vs Producer. UI vs Code.

Interactivity: Built-in reactive UI elements enable intuitive data exploration and visualization.

Reactivity: Cells automatically update when dependencies change, ensuring a smooth and efficient workflow.

Out of the box: Use sliders, textareas, buttons, images, dataframe GUIs, plotting, and other interactive elements to quickly iterate on ideas.

It's 'just' Python: Pure Python scripts for easy version control and AI coding.

• Reactive Execution: Run one cell, and marimo automatically updates all affected cells. This eliminates the need to manually manage notebook state.
• Interactive Elements: Provides reactive UI elements like dataframe GUIs and plots, making data exploration fast and intuitive.
• Python-First Design: Notebooks are pure Python scripts stored as .py files. They can be versioned with git, run as scripts, and imported into other Python code.
• Reproducible by Default: Deterministic execution order with no hidden state ensures consistent and reproducible results.
• Built for Collaboration: Git-friendly notebooks where small changes yield small diffs, facilitating collaboration.
• Developer-Friendly Features: Includes GitHub Copilot, autocomplete, hover tooltips, vim keybindings, code formatting, debugging panels, and extensive hotkeys.
• Seamless Transition to Production: Notebooks can be run as scripts or deployed as read-only web apps.
• Versatile Use Cases: Ideal for experimenting with data and models, building internal tools, communicating research, education, and creating interactive dashboards.

Advantages Over Jupyter Notebooks

• Reactive Notebook: Automatically updates dependent cells when code or values change, unlike Jupyter where cells must be manually re-executed.
• Pure Python Notebooks: Stored as .py files instead of JSON, making them easier to version control, lint, and integrate with Python tooling.
• No Hidden State: Deleting a cell removes its variables and updates affected cells, reducing errors from stale variables.
• Better Git Integration: Plain Python scripts result in smaller diffs and more manageable version control compared to Jupyter's JSON format.
• Import Symbols: Allows importing symbols from notebooks into other notebooks or Python files.
• Enhanced Interactivity: Built-in reactive UI elements provide a more interactive experience than standard Jupyter widgets.
• App Deployment: Notebooks can be served as web apps or exported to static HTML for easier sharing and deployment.
• Advanced Developer Tools: Features like code formatting, GitHub Copilot integration, and debugging panels enhance the development experience.
• Script Execution: Can be executed as standard Python scripts, facilitating integration into pipelines and scripts without additional tools.

Resources

• https://docs.astral.sh/uv/
• https://docs.marimo.io/index.html
• https://youtu.be/PcLkBkQujMI
• https://github.com/BuilderIO/gpt-crawler
• https://github.com/simonw/llm
• https://ollama.com/
• https://ai.meta.com/blog/llama-3-2-connect-2024-vision-edge-mobile-devices/
• https://qwenlm.github.io/