Concept B · order-first

Fulfillment

Finished vials on hand at each site. Every number here is the sum of recorded movements, never a typed-in total.

New order

On hand

Preparation	Texas · Elite Care	SF · Striker	Total	Paid fills	Shipped 30d	Counted	Cover
Semaglutide / B12 5 / 1 mg/mL · 3 mL	4	5	9	14	42	5 days ago	6 days
NAD+ 100 mg/mL · 5 mL	2	9	11	9	21	5 days ago	16 days
GHK-Cu 50 mg/mL · 5 mL	0	10	10	5	6	5 days ago	50 days
Tirzepatide / B6 5 / 2 mg/mL · 2 mL	3	3	6	12	36	5 days ago	5 days
BPC-157 3 mg/mL · 5 mL	0	1	1	11	28	5 days ago	1 day
10 more preparations, all zero at both sites

Sites are the two the code already knows about. Every count is the total of the movements recorded against it, so a number can always be traced to the events that produced it. There is no editable on-hand field to disagree with the ledger.

Days of cover is on hand divided by how fast the preparation actually left the shelf over the last 30 days. It is derived from our own shipping history, not from anything the supplier told us. It deliberately stops there: turning cover into a reorder date would need a supplier lead time, and we do not have one on record.

What is actually moving

last 30 days, from the ledger

BPC-157 — fastest-moving and nearly out
Only one vial remains, at SF, and its waiting list is already the third largest. Order BPC-157
Tirzepatide / B6 — second-fastest and nearly out
This is the one to order before it becomes the row above. Order Tirzepatide
GHK-Cu — slow-moving with the deepest cover
Overstocked relative to how it moves. Worth knowing before the next order repeats the quantity. No action

Deliberately not here yet: a reorder point, a suggested quantity, or a "you should order every N days". Those decisions need a buffer policy that nobody has set. What the ledger can honestly say today is how fast something moves and how long the shelf will last. How much to over-order stays a human judgement until that policy exists.

Change what is on hand

Add stock — vials arrived that are not against an order

The case this exists for: the supplier shipped extras by mistake and we kept them. They are real vials at a real site, so the count has to say so.

Site

Preparation

How many vials

Why

Note

SF BPC-157 goes 1 → 7. Nothing else changes.

Record a count — you counted the shelf and the number disagrees

Enter what you counted. The difference is recorded as the movement, so the ledger still explains every number.

Site

Preparation

Counted on the shelf

Note

Ledger says 9, you counted 8. Records −1 and resets the counted date.

Move stock between sites — same vials, different shelf

From

Preparation

How many vials

SF NAD+ 9 → 6 · Texas NAD+ 2 → 5

Movements

Every number above is the sum of these

When	Site	Preparation	Change	Reason	Because of	Who
2 days ago	SF · Striker	Semaglutide / B12	+5	Received	Order #12	Casey Review
2 days ago	SF · Striker	NAD+ 100 mg/mL	+2	Received	Order #12	Casey Review
2 days ago	SF · Striker	BPC-157	+1	Received	Order #12	Casey Review
4 days ago	SF · Striker	GHK-Cu 50 mg/mL	−1	Filled from stock	Rx #311 · Devon Review	Avery Review
6 days ago	Texas · Elite Care	NAD+ 100 mg/mL	−2	Transfer out	Paired with the entry below	Avery Review
6 days ago	SF · Striker	NAD+ 100 mg/mL	+2	Transfer in	Paired with the entry above	Avery Review
9 days ago	Texas · Elite Care	Tirzepatide / B6	−1	Adjustment	Vial damaged in handling	Casey Review
5 days ago	Both sites	All 15 preparations	set	Opening count	First count on record	Casey Review

What this page assumes

Stock is a balance, not a stage of work, so it sits after To ship rather than inside the pipeline. Shipping is what draws it down, so the causal order reads left to right.
On hand is derived, never typed. Every number is the sum of signed movements, matching the ledger PR #926 already implements. An on-hand column you can edit is a number that can disagree with its own history. "Opening count" is a movement like any other, and only one is permitted per preparation per site.
Where a received vial lands is still unsettled. 04-act records arrivals against individual patient fills while this page records the same arrivals as site stock. Both cannot be true for the same vial. PR #926 carries both a per-fill location and a stock ledger with a fill-from-stock path, so the model can express either; the choice is still open.
"Paid fills" is on this page on purpose. It is the only column that turns a reference table into a decision: 1 BPC-157 on hand against 11 waiting means order 10; 10 GHK-Cu against 5 waiting means order none. The same pairing appears on the To order tab, which is where the decision is actually made.
Transfers are two paired entries, not one move. Out from the sending site, in to the receiving one, sharing a note, so neither site's balance is ever briefly wrong and the pairing stays visible in the ledger.
Site names are the two the code knows about. The stored values are tx and sf; the labels here are the operator-facing names. Worth confirming the SF site's real name before it hardens — it has been called both "SF / Striker" and "CA / RonanRX HQ".
This whole tab needs PR #926's data model. PR #1059 has no stock model at all. The only quantity mutation in its codebase is a decrement on the internal-compounding branch, and inventory_lots has no site column. Nothing on this page can be built without PR #926's ledger.
Adding stock is a first-class action, not an edit. When the supplier ships extras we keep, that is a real event with a reason and a note, recorded like any other movement. There is still no editable on-hand field — the number moves because something happened to it.
Velocity comes free with the ledger. "Shipped 30d" is a sum over movements we already record, so knowing what is hot needs no new data. Days of cover follows from it. A reorder date or quantity does not follow until somebody sets a buffer policy.
Ordering starts from here too. Looking at what is low is the moment you decide to order, so the page offers it rather than making you navigate back to demand and remember what you just read.
All counts here are synthetic and consistent with data.md. They are not a report of real stock.