InfoPro
Keep the work whole.
InfoPro came from a simple pain: academic projects do not usually fail because one file is missing. They fail because the work gets scattered until nobody can see the whole thing anymore.
The pain was familiar.
A student would have a PDF here, screenshots there, sources in a chat, corrections in another chat, a supervisor comment buried somewhere, a draft saved as “final,” then another final after that. By the time the next change came, the first work was already hard to trust.
I saw how much time was wasted just rebuilding context. What topic are we on? Which version is current? Where did this source come from? What exactly changed? Why are we explaining the same thing again?
That is what “keep the work whole” means to me.
Not a slogan. A feeling. The frustration of watching serious work break into pieces before the real thinking is even finished.
So I started building around continuity.
The idea behind InfoPro is that a project should have memory. Research, drafting, revision and checking should not behave like separate strangers. They should work from the same project, the same document, the same sources and the same history of decisions.
That became the direction: a document-first academic workspace where chapters, sections, sources and revisions stay connected instead of floating around as disconnected tasks.
My part.
I founded InfoPro and lead the product and technical direction within PHANGS LTD. I set the direction around project memory, document-first workflow, review discipline and correction. I am not claiming I personally wrote every line or that every imagined capability is already finished.
What the system is trying to hold together
The document
The project is treated as one academic document with chapters, sections and revisions, not a pile of unrelated actions.
The sources
Research context is meant to stay connected to the writing, so later work does not have to start from a blank memory.
The corrections
Revision matters because academic work changes. The system has been shaped around making changes without pretending the earlier version never existed.
The boundaries
Project and document routes are owner-scoped. That is an engineering boundary, not a sweeping security promise.
It has not been a clean hero story.
Some paths worked. Some broke. Testing exposed places where the product could say or do the wrong thing: unreachable flows, citation problems, revision behavior that changed too much text, and export paths that were not ready to be promised.
I do not want to dress that up. The important part is that the failures changed the product. InfoPro moved more strongly toward one saved academic document, clearer project state, owner-bound routes, better revision behavior and a stricter standard for what is allowed to be called ready.
What is true now.
InfoPro is deployed and still being tested before I call every path ready. The product has real implementation behind the project-memory direction, but export stays out of the public promise until the exact live path works cleanly in the browser.
There was real academic-project work before InfoPro became the product, including paid manual work. That supports the origin story. I am not presenting it as InfoPro revenue, customer traction, market fit or proven user outcomes.
Where it is going.
The goal is simple to say and hard to build: when someone returns to their work, the project should still remember itself. The draft, the sources, the corrections, the reasoning and the next step should not be scattered across five places.
That is the product lesson I am carrying forward: useful AI is not just generation. It is memory, structure, correction and trust around the work.