Inspiration
Interruptions are common problems that many people face while they try to accomplish their work. While it usually only takes a very short pause to get distracted, it often takes a lot longer to remember what one had already done or what action one had already intended to take next. Our project aims to let a user rejoin their task without losing their train of thought.
What it does
Interrupt Me Later makes it possible to capture the current context of your work before you take a break. Users have to enter what they are doing at the moment, what has been completed so far, what problem they have just encountered, and what actions they are going to take next. The website processes this data and creates a user-friendly restart card. When users resume their work, they get a quick overview of:
- what has already been done
- what they have been currently working on
- what was blocking them from moving forward
- what actions they should take next.
Users can choose between two possible forms of the restart card: a simple one and a more detailed variation.
How we built it
We developed the website using Python and Streamlit. For the app, we used a rule-based analysis system that allows the website to read and analyze the user’s notes according to rules identifying the user’s completed tasks, obstacles or barriers, tasks to be done, and questions. The data we obtain is presented in the clear form of a restart card. Users can save their sessions in a JSON file to further go back to their previous checkpoints. To develop the project, we used Git and GitHub.
Challenges we ran into
Our primary difficulty was to configure the site that would interpret unstructured notes with no paid AI API. However, the users can use different phrases when describing their work, so we had to implement some rules that detect the typical phrases such as “finished,” “stuck,” “next,” and “not working.” The secondary challenge was to find the type of useful information that could be useful when resuming the task. We aimed at avoiding repetition of everything the user wrote down, therefore we had to find a balance between the amount of useful information and its reduction to the reasonable size of restart plan we wanted to create.
Accomplishments that we're proud of
We are proud that overall our website can take messy notes and automatically turn them into an organized checkpoint using only local Python code. We are also proud of the quick restart mode, which gives users only the most important steps they need to begin working again. The saved session history also makes the website feel like a complete tool rather than a one-time form.
What we learned
We gained an understanding of how to create and launch a Streamlit website more effectively, save user data in JSON format, and structure a Python project properly in different files. We realized that automation does not always require complicated AI technology. One can achieve automation using a simple rule-based solution and this can still solve a real-life problem. Most importantly is the need to come up with designs that take into account real issues.
What's next for Interrupt Me Later
We plan to incorporate a browser extension that will recognize open tabs, files being worked on, and tasks being performed before the user takes a break away from the computer. We would also like to add an option to use voice input, the cloud, and better storage options for longer periods of time and also integrate AI solutions for complex functioning.
Log in or sign up for Devpost to join the conversation.