We want to understand how research actually unfolds in real time — what strategies researchers use, what cognitive state they're in, and how that shapes the next steps. Our goal is to advance researcher-centered AI assistants: tools built to help researchers, not replace them. SciFlow is calling for your effort to do that: capturing your research in real time, through screen recording and your connected tools, from brainstorming to publication.
Questions about data ownership, what we will do with your data, or privacy? See the FAQ.
Below is what the collected data looks like for a six-month mechanical engineering project. The published result has a clean structure, but the real path has numerous dead ends and pivots that never make it into print.
We will install SciFlowApp, the app designed and developed by our team, on your laptop. It records your screen and connects to your workspace. It is also a data central hub: everything is transparent, and you can see all of your collected data in one place. You have full control over what to collect, what to exclude, and what to delete.
You choose exactly which folders, channels, repositories, projects, and screens are shared. Anything you don't approve is never collected — and you can revoke access to any stream at any time.
SW: re-run all three designs — Mar 24, 2:00 PM (from Google Calendar)
Set up SciFlowApp and choose which folders, channels, repos, and screens to approve.
Do your normally planned research; the tool records your approved streams in the background.
Review your recorded activity with us.
You do. You have full control over your data. Everything we do with your data will go through you.
No. We will not give your data to a big frontier lab or AI provider to develop AI models. The goal is to understand how real research unfolds so we can build AI to support researchers like you, not to replace them. If we ever wanted to use your data to develop a model, we would ask for your consent again.
Research is messy, and we understand that. But we believe that messy process is what's really valuable to learn from. It's also a chance to review and learn from your own workflow together — not a critique. You also have full control of what you choose to share with us: pause recording at any time, exclude any app or window from being captured, and delete any data points. For data release, you can opt out of releasing anything you don't want shared.
We account for that. For screen recording, you can select exactly which tabs or windows get recorded — your personal windows will not be recorded. The same goes for connected tools: only the specific Slack channels, Drive folders, and other items you approve are collected. And if there's ever anything you want deleted or don't want us to store, just tell us and we'll delete it.
You have total control over your data. You can opt out of releasing any individual data point you don't want shared. Beyond de-identifying your data, we'll triple-check with you about exactly what can and can't be released before anything goes out. From our side, we hope to contribute some of the collected data back to the research community, with your consent.
Your data will be treated as human subject data. Public access is gated and consent-based: anyone who wants to access it has to request it and get your consent, and it can only be used for research purposes — it is not freely distributed.
Beyond compensation (co-authorship, or monetary value), you can opt in to a monthly report on your workflow with the insights we find. And if we were to ever develop a personalized, researcher-centered AI, we'll prioritize your lab. Many participants also find the tool useful for keeping their research tools and progress in one place.
Reach out to the study coordinator about participating, or to the PI for general inquiries and cross-project collaboration.