Brief2Done：將作業要求轉化為清單的本地 AI 助手

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 3: Slow and Steady.

Brief2Done: I Built a Local AI Tool to Stop Procrastinating on Assignments

You know that feeling when your lecturer gives you an assignment, you open the instructions, see how lengthy or complicated they look, and immediately tell yourself, "I'll check it out tomorrow"?

Well, that's me sometimes. 😅

Tomorrow comes, and I postpone it again. Before I know it, the deadline is approaching, and I'm only just trying to understand what the lecturer actually wants.

Sometimes, the assignment itself isn't even the biggest problem. It's the effort of reading everything, identifying the requirements, and figuring out where to start.

So I began wondering:

What if I could build something that handles that first stressful part for me?

Something that could read an assignment brief, explain what I'm expected to do, and turn the important instructions into a simple checklist.

That's how Brief2Done was born.

For Hacktoberfest 2026's Slow and Steady challenge, I decided to turn that personal frustration into a working project powered by open-weight AI.

And yes, the goal is to be a little lazy—but in a productive way!

What I Built

Brief2Done is a local-first AI-powered assignment brief analyzer that helps students understand what their assignments require before they begin working.

Instead of spending time manually reorganizing lengthy instructions, students can paste an assignment brief into Brief2Done and get a clearer breakdown.

The application provides:

• Main task: A summary of what the lecturer is asking.
• Checklist: Important tasks and detected submission requirements.
• Word-count range: When a recognizable word range is included.
• Submission format: Such as PDF or DOCX.
• Deadline: When a weekday deadline is explicitly stated.
• AI suggestions: Helpful ideas for approaching the assignment.

For example, imagine receiving this assignment:

Compare supervised and unsupervised machine learning. Discuss at least three algorithms, provide two real-world applications, and include a comparison table. Submit your work as a PDF between 1,000 and 1,500 words by Wednesday.

Brief2Done can identify the main task and detect constraints such as the 1,000–1,500-word range, PDF format, and Wednesday deadline.

It's not designed to write assignments for students.

It's designed to make starting an assignment easier, so we have fewer excuses to keep postponing it.

Demo

Brief2Done runs locally through a simple Gradio interface.

The interface has two sections: one for pasting an assignment brief and another for displaying the analysis.

Here's Brief2Done in action:

**

**

I tested the prototype with different assignment briefs, including:

• An operating-systems report with a word-count range, PDF requirement, and deadline.
• A TCP versus UDP assignment without an explicit deadline or submission format.
• A machine-learning assignment requiring multiple algorithms, applications, a comparison table, and submission constraints.

One of my favorite moments during development was when I tested Brief2Done and saw it correctly identify the word count, PDF format, and deadline.

I was like, "Wow! So something that can help me understand what an assignment actually requires of me can exist like this?"

I was genuinely amazed.