Skip to content

Every jar remembers.

Give a plate, a slant, a grain jar or a bulk block one five-character name. Scan it, say what you did, and the lineage and the numbers build themselves.

Invite-only while the first labs run it.

Every day01

The lab forgets faster than you do.

You run a clean hood and good tek. Then a tray comes back green, and the questions start. Which grain batch? Which sterilizer run? Which transfer, off which plate, on which day?

The jar can't tell you. The Sharpie ran in the pressure cooker. The spreadsheet stopped four weeks ago, on the week you got busy, which is exactly the week you'd want to read back.

So you guess. You throw out a line that was probably fine, or you keep one that isn't. Both cost a season.

A lid you can't readTwo jars, one labelSpreadsheet · last edited 34 days agoA strain that used to run fast
Wrote it before the pressure cooker. Steam won.
Last edited 34 days ago. During the busy month.

K7F2Q · Grain jar · G3 · HE-Aggressive-03 · Incubator 1

One day02

Five characters change everything.

A vessel gets a code. Five characters from a 32-character alphabet with I, L, O and U taken out, so a 1 is never an I and a 0 is never an O on a steamed-up lid. There are 33.5 million of them.

It goes on twice. You write it by hand before the jar goes in the pressure cooker, because handwriting survives steam. You stick a printed QR label on it at the hood, because a sticker survives handling.

From then on it isn't “the third one on shelf B.” It's K7F2Q. Point a camera at it and its whole life is on screen.

Write a 5 that reads like an S? The app knows which characters people confuse by hand, finds the vessel you meant, and asks you to confirm before it writes anything.

Because of that03

One name becomes a family tree.

A transfer takes under ten seconds. Pick the child type and how many. The children inherit the strain, step the generation up by one, and get their own codes to label.

Nobody types a lineage. It accumulates.

Months later a bulk block goes down with trichoderma, and you can walk backwards from it to the grain jar, the liquid culture, the plate and the original spore print. You can also walk forwards, to every sibling that came off the same source, so you know what else to check tonight.

Generations count from G0. Past about G7 vigour tends to fall away, so the app flags a line when it's time to go back to master stock.

A lineage tree from one spore print down through five generations.One bulk block at generation four is contaminated. The path back through its grain jar, liquid culture and original plate is highlighted, and the two vessels that came off the same liquid culture are flagged for checking.Trichoderma · totalG0G1G2G3G4

Both of these came off the same liquid culture. They are what you check tonight.

Because of that04

Say it once, with gloves on.

The best record is the one you actually make. At the hood, with wet gloves and a flame going, a form is a form. So talk instead.

  • “Transferred K7F2Q onto four MEA plates from batch twelve, shelf B.”

    Four plates created, coded, parented, generation stepped, media batch attached, location set.

  • “Log 80% colonization on 4T9RK, rhizomorphic, looking clean.”

    One observation, timestamped, on the right vessel.

  • “Which strain colonizes grain fastest?”

    A straight answer, computed from your own vessels.

It proposes, you confirm. Nothing destructive happens on its word. Marking a vessel contaminated or retiring one waits for your yes, and it reads the code back before it writes.

The agent runs on the same tool surface as the screens, so it can't do anything you couldn't do by hand, and you can't do anything it can't see.

Because of that05

Now the lab has an opinion.

Nothing on this screen is typed in. Every number is computed from work you already logged.

Average days to colonize, per strain. Contamination rate, sliced by media batch, sterilizer run, operator, vessel type, month or line. Biological efficiency — fresh yield over dry substrate weight — per block and averaged per strain.

That's how a hunch becomes a decision. “This isolate feels faster” turns into 11.4 days against 16.2. “Something's off with the grain” turns into one batch at 31% and the rest at 4%.

HE-Aggressive-03

Hericium erinaceus

12

active

11.4d
Avg days
6%
Contam
94%
BE%
28%
27%
16%
12%
8%
4%
MarAprMayJunJulAug

Bar height is vessels started that month. Red is the share that failed.

Until finally06

You stop guessing.

A year in, the shelf looks the same. What changed is that you can ask it things.

Which line to run at scale. Which one to retire. Whether the new sterilizer cycle actually helped. Where the contamination gets in, and on which step.

You didn't become a different cultivator. You just stopped throwing away what you learned.

Three things, then it runs itself.

  1. 01

    Label

    Print a sheet of QR codes, or write five characters on the lid. A code becomes real the moment you claim it.

  2. 02

    Scan

    Camera at the label, or type the five characters. You land on that vessel with every action one tap away.

  3. 03

    Say what you did

    Observe, transfer, move, harvest, contam, retire. Tap it or say it. Everything else is computed.

Built for a lab you can walk across.

The hobbyist who outgrew the notebook

You have more than twenty active vessels, two or three lines you actually care about, and you've started losing track of which plate came off which. The notebook was fine at ten. It isn't at fifty.

The small commercial farm

You sell what you grow, so contamination is money and consistency is the product. You need to know which strain goes on the shelf next month, and you need to trace a bad block back to the batch that made it.

The research or teaching lab

Several people touch the same cultures. Provenance has to survive a change of hands, a semester break, and a student who graduated. Every write is timestamped into an audit log with the raw input kept.

It isn't for everyone. Multi-user roles, environmental sensors, inventory and sales, sequencing data, offline sync: all deliberately out of scope right now. If those are the reason you're shopping, this won't be it yet.

What you're probably using now.

Memory
Excellent for a week. The problem is that the expensive question is always the ten-week-old one, and that's exactly the one memory can't answer. Nothing to install and nothing to maintain, which is why everyone starts here.
A paper notebook
Reliable, cheap, survives a dead battery, and a good one is genuinely better than bad software. But you can't search it one-handed at the hood, and it can't average anything. Ten pages of careful observations still won't tell you your contamination rate by grain batch.
A spreadsheet
It computes, which puts it ahead of the notebook. The catch is upkeep: a spreadsheet wants a keyboard, later, from memory, so it decays exactly when the lab gets busy. And a row has no lineage. You can put a parent code in a column, but nothing walks the tree for you.
Inventory or ERP software
Built for stock that arrives, sits and ships. A culture isn't stock. It divides, it steps a generation, it inherits genetics, and its real value is a performance history. There's nowhere in those systems to put any of that.

This app is narrow on purpose. It keeps the identity and history of living cultures, on a phone, in the seconds you actually have.

Start with one jar.

Write five characters on a lid. Scan it. That's the whole habit. Everything on this page grows out of it.

Myco Lab Tracker is invite-only while the first labs run it. Tell us about your lab and we'll come back to you.