You labeled every tube. Nobody labeled the folders.
BenchTrail reads your lab’s Google Drive and rebuilds each experiment as a dated timeline, so the next person (or you, in four months) can find out what happened.
Free during the pilot. Email us your name, your field and the Google account your lab’s Drive is on, and we’ll set you up. Read-only access to Drive, and nobody writes anything up first.
- Protocol
- Imaging, round 1All 24 cells imaged.
- Imaging, round 2Cell_7 has no round 2 files. Why?
- Spacing analysis, rerunNew mask threshold. Median 190 nm.
- Group meeting
Everything’s there. Finding it takes a morning.
The protocol is in one folder, the raw data in another, and the analysis wherever Claude happened to save it. The person who could walk you through it may have left the lab.
One experiment, laid out on the bench
Spacing shows the real gaps between steps. Pick any strip to see what BenchTrail found behind it.
- Protocol
- Imaging
- Analysis
- Slides
Mar 14, 2026
Imaging, round 2
Imaged the same cells again for the second round. 23 of 24 have files.
Cell_7 has no round 2 files. Dropped on purpose, or are they somewhere else?
- Cell_1_R2_000.ome.tif
- Cell_2_R2_000.ome.tif
- +734 more, listed by name
Example data. The experiment and the people in it are made up.
What your lab does with it
-
to check
Answer the things only you know
BenchTrail asks about gaps in the files, like the missing Cell_7. You tap an answer and it’s saved beside the step with your name on it. Anything AI wrote stays marked as a draft until someone in the lab confirms it.
-
why rerun??
Ask what happened, and get the file
Ask in plain English, in BenchTrail or from Claude. Every answer links to the step and the file it came from, and it says so when the record doesn’t know.
-
last day: Fri
Hand over more than a Drive link
When someone leaves, make a dated handoff pack: what was done, what’s still open, where the data lives, and what could get lost, like imaging data sitting in a personal My Drive. It works with AI off.
DO NOT TOUCH
That’s the rule BenchTrail runs under. It reads your Drive and leaves everything exactly where it found it.
- Read-only, by construction
- Access uses Google’s
drive.readonlyscope, and the Drive client can only make read requests. - Raw images stay put
- Image files are listed by name and never read. Text is read from protocols, notes, scripts, documents, spreadsheets and slides.
- AI has an off switch
- Turn AI off for the whole lab, and timelines still build from your files.
- You can see what was sent
- Every Drive request and everything sent to an AI model is recorded in an audit log, by lab and member.
- Only your lab can see it
- Records are scoped to your lab, and each build sees only what the person who started it can see in Drive.
- What it keeps
- A private copy of the text files it reads, so search and rescans work. Email us and we’ll delete it.
I started BenchTrail after watching my brother, a microscopy researcher, retrace his own analyses through Drive from memory, one directory at a time. The protocols, raw data and slides all existed. They just lived in different places, and only he knew where.
If that sounds like your lab, I’d like to build your first timeline with you.
RahulRahul Roy, founder · rahul@getbenchtrail.com
What labs ask first
Do we have to rename files or change how we work?
No. BenchTrail works from the folders you already have. It does need an experiment ID or name that shows up somewhere in the folder names, file names or file contents, because that’s how it links files to an experiment.
What does the pilot cost?
Nothing, and you don’t need a card to start.
Why do I have to request a spot?
Google hasn’t finished verifying BenchTrail yet, so until it does, only Google accounts we add by hand can connect Drive. Send us the account your lab’s Drive is on and we’ll add it, usually the same day. When you connect, Google will say the app is unverified or in testing. That’s expected during the pilot. Drive access also expires every 7 days for now, and BenchTrail will ask you to reconnect.
Does this replace our notebook or Benchling?
No, it sits beside them. Today it reads Google Drive only, so anything that lives solely in another system won’t appear.
Is it only for microscopy?
It was built with a microscopy lab, so imaging rounds, per-sample folders and repeated analysis passes are what it understands best. Other experiments still get a timeline from their dated files and folders.
What goes to AI?
With AI on, file names and text excerpts are sent to Anthropic’s Claude to draft step descriptions and answer questions. An optional file-ranking step uses TypeSafe’s Jev. Raw images are never sent. A lab admin can turn AI off at any time.
Who can see our experiments?
Members of your lab. Each member connects their own Google account, and a build sees only what that account can see.
Pick one you’d struggle to explain from memory, and connect its Drive folders to see its timeline. The pilot is free, and we add labs to it by hand.
Request a pilot spot