ActionInbox

Inspiration

People receive important emails every day but often do not know what requires action, what can wait, or what deadline they may miss.

What it does

ActionInbox turns emails into clear, evidence-backed tasks.

It identifies messages that require action, extracts deadlines and important details, and explains what the user should do next.

Every extracted task is connected to evidence from the original email, helping the user verify that the AI did not invent the information.

How we built it

ActionInbox is being built as a web application using Python, FastAPI, SQLite, the Gmail API, OpenAI GPT-5.6, Docker, GitHub Actions, HTML, CSS, and JavaScript.

Challenges we ran into

The main challenges are reducing AI hallucinations, safely handling untrusted email content, and creating a simple experience without allowing automatic or dangerous actions.

Accomplishments that we're proud of

We are building an end-to-end workflow that transforms a long email into clear tasks, deadlines, supporting evidence, and personalized guidance.

What we learned

We learned how to combine AI analysis with structured output, source evidence, security boundaries, and a practical DevOps workflow.

What's next for ActionInbox

Future versions may support business knowledge bases, team task assignment, Outlook, calendar suggestions, and additional communication platforms.

Built With

• css
• docker
• fastapi
• github
• gmail-api
• gpt-5.6
• html
• javascript
• openai-api
• python
• sqlalchemy
• sqlite