Eternal Gardens for research groups

Carry your lab's knowledge forward

Every research group builds up workflows, tools, and hard-won know-how over the years. As members move on to new chapters, EG helps that knowledge carry forward — gathering it into an interactive Mentor your team can simply ask, so each new student can build on what came before instead of starting from scratch.

Open my lab See plans

What it's for

A research group's most valuable asset is often undocumented: the exact order of steps, the parameter that has to be converged first, the script that looks deprecated but isn't, the reason a previous approach was abandoned. It lives in one person's head. When they graduate, it's gone.

EG Knowledge Transfer turns that tacit expertise into a Mentor — an AI grounded in the files a member uploads and a short guided interview that draws out what the files can't. Current and incoming members ask the Mentor questions in plain language, follow its handoff guidance, and take quick quizzes to check they've understood. Nothing walks out the door.

Who does what

Supervisor / PI

Owns the lab. Creates projects, invites the team, appoints Group Admins, and watches knowledge get captured.

Group Admin

A senior student or postdoc who runs day-to-day capture: creates projects and Mentors, curates sources, builds quizzes.

Departing member

The person whose knowledge you're preserving. Uploads their files and answers a short interview before they leave.

Incoming student

Asks the lab's Mentors how to do things, browses the sources behind each answer, and takes onboarding quizzes.

Set it up in five steps

Everything below lives in Account → Academic Lab. The Start here tab tracks your progress as you go.

1

Create a project

A project is a body of work — a technique, an instrument, a codebase. It groups the Mentors and source files that belong together, and scopes questions and quizzes so answers never mix across unrelated work.
2

Create a Mentor for a departing member

Give it their name or the workflow it captures, assign it to a project, and note the domain. That's it — no avatars or backstory required. A Mentor is a container for one person's or one workflow's knowledge.
3

Add their sources and run the interview

Upload scripts, notes, input files, slides, and data. Then answer the short Knowledge interview — it asks about setup, assumptions, failure modes, validation checks, and "what should a new student never miss?" This is where the tacit know-how gets recorded. Sensitive input files can be uploaded as confidential so they're encrypted and never sent to any AI provider.
4

Invite your team

Add students and co-supervisors by email from the Members tab. They get access to the lab's Mentors and projects; Group Admins can help you run capture.
5

Ask, and check understanding

From the Ask tab, anyone can ask a Mentor how to do something and get a grounded, cited answer with links to the exact sources. Use Quizzes to give incoming students a quick handoff check — and to see where the gaps are.

Common questions

What exactly is a "Mentor"?

An AI grounded in one member's uploaded files plus a short interview. It answers questions about their workflow in plain language and points to the source behind every answer. It is not a chatbot personality — it's a knowledge capsule.

Is our data private?

Yes. A lab's Mentors and sources are visible only to active members of that lab. Files marked confidential are encrypted on our server and are never sent to any external AI service.

How long does capture take?

A departing member can upload their files and complete the core interview in an afternoon. You can always add more later; readiness improves as you do.

Who can create and edit Mentors?

Supervisors and Group Admins can curate every Mentor in the lab. Any member can create and train their own. Everyone with a seat can ask questions and take quizzes.

Preserve what your lab knows

Start with one project and one departing member — you'll have a working Mentor the same day.

Open my lab