Inspiration
Have you ever opened an old project and realized you no longer remember why you made certain decisions?
Maybe you tried several approaches before choosing one. Maybe you abandoned an idea that could actually work today. The files are still there, but the reasoning behind them is buried in old notes, documents, and previous versions.
I wanted to create something that could help recover that lost context, instead of forcing people to piece everything together manually.
What it does
Project Archaeologist is an AI assistant designed to reconstruct the history of a project from files stored on a personal NAS.
It helps answer questions like:
- Why did we choose this approach?
- What changed between earlier and newer versions?
- Which ideas did we abandon, and why?
- What was left unfinished?
- Is there anything from an older version worth bringing back?
The idea is to go beyond searching and summarizing files. Project Archaeologist connects information across documents, creates a timeline of important changes, and helps users rediscover forgotten decisions and ideas.
It also aims to show where its answers come from. If there isn't enough evidence to explain a decision, it should say so rather than make something up.
How we built it
The project is being designed as an AI Skill for UGREEN NAS.
The planned workflow starts by finding relevant project documents, notes, and saved versions. The AI then extracts important changes and decisions, connects related information, and organizes the results into a timeline.
Users will be able to ask questions about the project's history and receive answers supported by their original files.
The initial version will focus on documents and saved file versions available through UGREEN's supported tools.
Challenges we ran into
One of the biggest challenges is figuring out the difference between what actually happened and what AI assumes happened.
For example, two versions of a document might show that something changed, but they don't necessarily explain why.
Another challenge is connecting information spread across different files without losing context.
We're also looking into which file metadata and version-related features are available through UGREEN NAS, so the project can work within the platform's limitations.
Accomplishments that we're proud of
The most important part of the concept is its focus on recovering the reasoning behind past decisions, rather than simply generating summaries.
We've outlined three main features:
- Decision Timeline: reconstructs important changes and decisions.
- Why Engine: searches for evidence explaining why a decision was made.
- Resume Mode: helps users return to an unfinished project and understand where to continue.
A key design principle is that the assistant should clearly distinguish documented facts from uncertain conclusions.
What we learned
Working on this idea made us think differently about personal file storage.
A collection of documents can tell a much bigger story than any individual file. Old drafts, notes, and versions can reveal how a project evolved, what problems appeared, and which ideas were never fully explored.
We also realized how important it is for AI systems to acknowledge uncertainty, especially when trying to reconstruct past events.
What's next for Project Archaeologist
The next step is to implement and test the Skill using UGREEN's supported tools.
We want to validate how accurately it can reconstruct project timelines, identify decisions, and connect information across different documents.
In the future, we'd like to explore Git integration, richer visual timelines, and ways to make the assistant useful for entire teams.
The long-term goal is simple: make it easier to return to any project, understand its history, and continue working without starting from scratch.
Built With
- ai
- markdown
- nas
- skills
- ugos
- ugreen