Spikelet

Spikelet turns a seed lot's own numbers into a drill setting, entirely on your phone. No account, no signal, no subscription.

WHAT IT ANSWERS
1. What does the drill have to be set to for THIS lot, on THIS field?
2. Is that germination figure even usable yet?
3. What is the drill actually delivering, as against what it was set to?

WHY BAGS PER ACRE IS THE WRONG UNIT
- A seeding rate is derived, not habitual. Target plant stand, divided by PURE LIVE SEED (germination times purity) and by field establishment, times the lot's own thousand-seed weight.
- Thousand-seed weight dominates. The same variety at 52 g needs over a third more seed by weight than at 38 g for the identical plant stand - so one habitual rate is right for exactly one lot and wrong for every other.
- Pure live seed does the same job at the other end. A 92% germination at 98% purity is 90% pure live seed, and the 10% that is neither has to be replaced or the stand comes up short.
- Establishment belongs to the FIELD, not the seed. A cloddy October seedbed and a fine spring one are different fields even when they are the same field, and the app keeps a separate figure for each.

A GERMINATION TEST IS NOT A RESULT UNTIL IT IS FINISHED
- Every species has a prescribed final count day - ryegrass runs fourteen. Read a tray early and it reads low, because the slow half has not finished. Spikelet reports that as an UNFINISHED test with the days remaining, never as a low germination.
- Four replicates that disagree wildly are a badly set test, not a variable lot. The app carries the tolerance table, shows the spread against it, and refuses to certify a test that exceeds it rather than averaging it.
- Normal seedlings only. A seedling with no root is not a plant in a field, and counting it inflates the figure the drill is then set from.
- A cold test, when you run one, overrides the warm germination - because that is what field emergence actually tracks.

IT REFUSES RATHER THAN GUESSES
No thousand-seed weight on the label? No rate. What you get instead is the exact kilograms per hectare PER GRAM of TSW, plus the plausible range for that species - so the one missing figure is the only thing outstanding, and weighing a thousand seeds finishes the job.
A calibration run too short to weigh? Refused, with the distance that WOULD put a usable mass on the scale. An error of three percent found by a five percent measurement is worse than no answer.

WHAT IS IN IT
- The derivation shown as a column you can read top to bottom and tap to change
- Seed lot register with lot reference, variety, supplier, purity, TSW, tags and archive
- Germination bench: replicate counters, tolerance band drawn, count schedule per species
- Fields with their own target stand and establishment, and the seed each one needs
- Drill calibration: what the run should collect, what it did, and whether the error is real
- Seed spacing down the row, so you can check the drill with a tape and no scales
- A shipped reference of 27 species: count days, TSW ranges, stands, establishment, row widths
- Insights, including rate against thousand-seed weight with your own lots plotted on it
- Export to JSON and CSV - including every replicate count, so a test can be re-checked
- Hectares or acres, metres or feet, light and dark themes, four accents

PRIVACY
No account. No ads. No analytics. No third-party SDKs. No data collected and none shared. The app requests no runtime permissions at all. Its single network request is loading the privacy policy in Settings; everything else works in flight mode.

The species reference is a published guide, not a warranty. Your own seedbed, your own drilling date and your own laboratory beat any table.