Meeting Notes to Action Items

Turn rough meeting notes into a clean summary, action items, risks, and decisions.

This is a small AI workflow project built for a professional portfolio. The goal is to show a practical business use case, a polished interface, and a clear example of structured LLM output without requiring a large or complicated codebase.

Project overview

This app takes unstructured meeting notes and turns them into:

• A concise summary
• A list of action items
• A list of risks or blockers
• A list of decisions made during the meeting

It supports two modes:

• Demo AI: local rule-based parsing so the app works immediately
• Live AI: OpenAI-powered structured output using the Responses API

Why this project is useful

• It solves a real problem that businesses understand immediately
• It demonstrates prompt design and structured outputs
• It shows how AI can turn messy input into organized workflow artifacts
• It is simple enough to run locally, review quickly, and discuss in interviews

Concepts this project explores

• LLM-powered workflow automation
• Structured output generation with JSON schema
• Converting messy text into usable business deliverables
• Clear fallback behavior when an API key is not present
• Separating frontend display logic from backend AI orchestration

Tech stack

• Node.js
• Vanilla HTML, CSS, and JavaScript
• OpenAI Responses API
• Local .env configuration

Project structure

meeting-notes-to-action-items/
  data/
    sample-notes.txt
  public/
    app.js
    index.html
    styles.css
  .env.example
  .gitignore
  package.json
  README.md
  server.js

How to run the project

1. Clone the repository

git clone <repo-url>
cd meeting-notes-to-action-items

2. Start the server

node server.js

3. Open the app in your browser

Visit:

http://127.0.0.1:3000

4. Try the app

• Click Load Sample to insert example meeting notes
• Click Analyze Notes to generate results
• Or paste your own notes into the text area

Demo mode vs Live AI mode

Demo mode

If no OpenAI API key is configured, the app still works using a built-in demo parser.

This is useful because:

• the project is runnable right away
• reviewers can test the UX without extra setup
• the app still demonstrates the intended workflow

In demo mode, the badge in the output panel shows Demo AI.

Live AI mode

If an OpenAI API key is configured, the app sends the notes to the OpenAI Responses API and requests structured output using a JSON schema.

In live mode, the badge in the output panel shows Live AI.

How to enable live OpenAI output

1. Create a local .env file

In the project folder, run:

cp .env.example .env

2. Add your OpenAI API key

Open .env and update it to look like this:

OPENAI_API_KEY=your_real_api_key_here
OPENAI_MODEL=gpt-5.4-mini
PORT=3000

Notes:

• Do not leave your_api_key_here in place
• Do not add spaces around the =
• .env is ignored by Git and should not be committed

3. Restart the server

If the server is already running:

• Press Control + C in Terminal
• Run node server.js again

4. Analyze notes again

When live mode is working correctly, the app should use OpenAI and show Live AI in the mode badge.

How someone else can use this repo

A visitor can use this repository in two ways:

Review the code and project structure

This is useful for employers, clients, or collaborators who want to understand:

• the problem being solved
• the interface and user flow
• the backend request logic
• how structured AI output is handled

Run the project locally

Anyone can clone the repo and run it:

git clone <repo-url>
cd meeting-notes-to-action-items
node server.js

If they want live AI output, they also need their own OpenAI API key in a local .env file.

Key files

• public/index.html: page structure and app layout
• public/styles.css: visual design and responsive layout
• public/app.js: frontend behavior and result rendering
• server.js: static server, demo parser, and OpenAI request logic
• data/sample-notes.txt: example notes for testing

Troubleshooting

The app says the API key is incorrect

Check that .env contains your real OpenAI key, not the placeholder value from .env.example.

The app is still using demo mode

Make sure:

• .env exists
• OPENAI_API_KEY is set
• you restarted the server after editing .env

I cannot find .env

.env is a hidden file. You can create it in Terminal with:

cp .env.example .env
open -e .env

The server will not stop

Press Control + C in the Terminal window where node server.js is running.

Future improvements

• Add export to Markdown or email-ready format
• Add copy-to-clipboard buttons for generated sections
• Add upload support for transcript files
• Add saved history for prior meeting analyses
• Add rate limiting and authentication if the project is later deployed

Portfolio talking points

If you discuss this project with a client or employer, a few honest talking points are:

• “I built a tool that turns messy meeting notes into structured follow-up outputs.”
• “I used OpenAI’s Responses API with structured JSON output.”
• “I designed the app so it still works in demo mode without requiring API setup.”
• “I focused on a practical business workflow rather than a novelty AI demo.”