MenuClose
Applied Sciences DivisionResearch → Prototype → Proof

DRKNYT / Applied Sciences / TAGD products & production

Qualify the line. Then repeat it.

Commission the label, invitation, packaging and 3D stations, then release measured batches with traceable results.

Division edition 04 Updated October 30, 2026

Inside the imagined lab · Image details

Generated garage-lab concept, October 30, 2026. OpenAI image tool, directed in Codex by LUCIEN. This is an imagined research environment, not a photograph of an existing DRKNYT facility. Image-model version was not exposed.

Bottle ID: fit, code and

The Bottle ID and the 48-label moving kit need different tests. For the top-mounted ring, start with the exact bottle and removable identification ring, top/side/underside photographs, caliper measurements and the actual printer available. Access to a -capable filament printer with a heated bed, /slicer, calipers and basic finishing is enough to start fit trials. A small outsourced prototype run is another option; neither route is a purchase or production qualification.

  • Record model and specimen, ring dimensions, underside attachment, and spout/hinge/latch clearances. Build an adjustable master and print unpersonalized fit samples first. Do not replace the factory lid, seal or drinking components.
  • Qualify one documented filament, machine, nozzle, orientation and slicer profile. Use for functional trials; use for shape checks and only if a flexible feature is justified. A copied molded snap is not automatically suitable for layered printing.
  • Establish a complete readable code area before finalizing decoration. Use a separate protected printed face for the first hybrid trial, with a genuinely encoded bottle-specific target and preserved clear margin. Keep the owner's referral identity separate.
  • On the finished attached part, record installation/removal, retention, opening/closing clearance, edges, cleaning, graphic adhesion and scans on the intended phones. Repeat after documented wear. Record failures and corrections; do not infer food-contact or dishwasher suitability from the filament name.
  • Retain the accepted and record consumed mass, machine time, hands-on personalization/assembly, waste and rejects. Then assess a 10-unit trial; no physical pass or batch capacity is asserted here. Assess product safety obligations for the intended age group before sale.

Files to produce for the ring

DeliverablePurpose and status
Editable master and exportMeasured geometry, attachment and tolerances; still to produce.
Printer-specific projectPart arrangement and selected slicer/material settings; still to qualify.
exports with documented unitsGeometry-only handoff; not the editable master.
Editable vector artwork (SVG)Approved mark, name/number and genuinely encoded code; mockup patterns are not usable code masters.
Printer-specific G-codeGenerated from the qualified project for the exact machine/nozzle/material; not a reusable master.
Batch record: spreadsheet (CSV) or structured data file (JSON)Links order/personalization, color, part revision and the correct unique object code. Keep personal and production allocation records private.
Inspection and cost recordLinks each finished part to fit/scan results, material lot, timing, rejects and release disposition.

These are proposed manufacturing deliverables, not files claimed to exist or download here. A correctly printed part with the wrong person's code fails inspection.

Relay: a proposed 25-token story study

This study is separate from Bottle ID fit trials and the 48-label moving kit. It combines the feel of the object with a consent-based story and handoff. No tokens or live chapters are created by this checklist.

  • Choose one physical format. Compare roughly two-inch prototypes for thickness, edge comfort, handling and finish before choosing a material/process. Use printed bodies to explore form where useful; establish a separate process and quote if metal is selected. The cyan accent does not require an illuminated part. Envelope, opening card and any optional presentation box need their own packed dimensions and assembly trial.
  • Prove one finished code first. Produce actual encoded artwork for one token, retain its fixed object identity and compare every decoded result to the private production record. Test the selected finished face at actual size, on intended phones and after recorded handling/wear. A printed mockup or metal-like rendering does not establish scan performance. Stop if the Earth treatment cannot be read reliably; revise and re-test before making the edition.
  • Define the story and handoff actions. The originator selects an opening premise. Test permitted reading, choosing to add a chapter, choosing its audience, and separately offering/accepting a handoff. A scan must not create ownership, publish or transfer anything. Receiving a token must not alter earlier authors' permissions. Define an unaccepted, declined, lost or interrupted handoff and how an authorized person recovers from it.
  • Test consent and continuity. Cover author correction/removal, private versus shared content and what survives a handoff. Keep private personal details and precise location out of public defaults. Define hosting responsibility/duration and export/closure handling before inviting people. If notifications are included, require explicit opt-in and a deliberately published new chapter; a scan is not the notification event.
  • Then assess the numbered edition. Map each proposed number, physical token and unique code to the correct story, recording accepted/rework/void units without silently reusing another token's identity. The target is 25 accepted tokens, not 25 presumed-good prints; stock, time, rejects and costs still need measurement. Do not infer costs or throughput from the Bottle ID or generic-part planner.
  • Observe a willing handoff. Record what people understand, whether they want to give the object away, whether a consenting next holder can read and participate, and where the experience fails. No rewards, leaderboard, map or complex mission is needed for this first test. Permission failures, wrong-code mapping and unresolved continuity obligations hold distribution.

TAGD owns the story, consent, identity and handoff implementation/qualification; Applied Sciences owns physical form, packaging, manufacturing records and finished-code tests. The experiment remains planned until the physical and digital checks pass.

Moving-kit calibration packet

Prepare one packet with 48 labels on two Avery 22805 sheets, a starter card and a flat sleeve. This moving-kit calibration packet is distinct from the invitation and its mixed sample assortment. This is the calibration sample for the production line. After it passes, time a representative batch and establish a repeatable work instruction.

The prepared digital layout has passed all-48 decoding and template-geometry checks. Physical printing, adhesion, live claiming and household transfer have not been performed. The saved test files and manifest are retained for the internal pilot and are not available from this page. Candidate allocations and printable production-code files are kept out of this public report.

1. Confirm the printer, paper and settings

  • Confirm the physical printer and its ink, condition and ability to feed the selected stock. The saved test packet was prepared for an existing HP ENVY queue. The production candidate needs its own permitted tray/media procedure; the old queue is not qualification for the new machine.
  • Confirm the package is Avery 22805, US Letter, 24 square 1.5-inch labels per sheet. The final uses the rectangles from Avery's downloaded official blank template; a similar-size label product may have different spacing.
  • Select standard US Letter, single-sided / Duplex None, Actual Size / 100%. Disable Fit to Page, Scale to Fit and borderless enlargement. The earlier development queue defaulted to duplex; explicitly verify single-sided output on the selected queue. These instructions do not change the driver automatically.
  • Select the media/feed settings appropriate to the actual stock and printer. Record the queue, driver, tray, media type and quality settings.

2. Calibrate on plain paper before using label stock

  • Print both saved label pages on plain paper at the settings above. Overlay each on a blank physical label sheet and check the printed square boundaries against the die cuts.
  • Measure horizontal and vertical displacement at the four corners and a central label on each sheet: ten positions total. Proposed acceptance is no more than 0.03 inch (0.76 mm) displacement in either direction at every position, with no clipped code, quiet zone or text. Record the values before declaring pass.
  • Stop on skew, scaling error, clipping or feed failure. Correct the relevant setting and repeat the plain-paper check; record wasted sheets. Do not compensate by stretching the image.

3. Print and inspect one complete packet

  • Print the two label pages single-sided on two stock sheets. Record all sheets used and rejected.
  • Confirm 24 labels per sheet, 48 total, with the human-readable codes matching the saved manifest. Confirm no code or text lies outside its label, and no code is stretched or its surrounding blank margin clipped.
  • Keep the 48 printed candidates together. They were generated offline, pass the source code format/checksum contract and are unique within this packet. This does not establish global allocation uniqueness, a registered database entry or a functioning live route.

4. Perform physical optical checks without opening live targets

  • Sample positions 1, 10 and 24 on each sheet: six physical labels covering a corner, the central region and the opposite corner.
  • On two phones, decode every sampled label in two recorded lighting conditions: well-lit indoor light and the ordinary light where the labels will be used. This is 6 labels × 2 phones × 2 conditions = 24 attempts. Record device, operating system, decoder/app and actual conditions; do not claim a measured light level without a meter.
  • Use a decoder that displays the encoded text without fetching it. Do not open, tap through, navigate to or automatically preview the encoded production URL during optical calibration. Compare every result to that label's exact manifest target.
  • Proposed acceptance: all 24 attempts return the exact target on the first attempt. Record retries and failures rather than omitting them. Any mismatch stops the run for investigation. This is an optical test, not a sign-in, claim, privacy or placement test.

5. Check adhesion on spare blank labels

  • Preserve the 48-code packet. Use three sacrificial blank labels from remaining stock on a clean, dry corrugated-cardboard test panel. Record the surface and conditions.
  • Apply each with ten firm fingertip presses. Inspect at 15 minutes and 24 hours. After the 24-hour inspection, make ten dry-finger rubs across each label and inspect all edges and corners.
  • Proposed acceptance: no edge/corner lift greater than 1 mm after these checks. Record observed lift, damage and conditions. These proposed thresholds are an initial screen, not an established durability standard or a claim about moisture, heat or long-term use.

6. Pack, measure and decide

  • Print and trim the starter card. Place the two sheets in sheet order with the card in a sleeve or envelope that holds Letter sheets flat. Do not fold the sheets or let sleeve adhesive touch labels. Keep the manifest separate; do not print household details on the sleeve.
  • Record preparation, printing, inspection, assembly and packing time separately; record used/spoiled stock, ink observations, packing cost, freight and any unmeasured cost.
  • Stop on feed, registration, scan, adhesion, code/manifest or material cost failure. Preserve the failed sample and record the correction before making more packets.
  • Accept physical preparation only when its tests pass. Release a representative repeat batch only with a recorded result and the TAGD team’s qualified household workflow. First authenticated claim and verification of the exact placed code remain required; the current owner runbook does not establish working authenticated persistence.

Product workflow before distribution

The square-stock exporter still needs correction: selecting 24-up square stock routes through 30-up address-label geometry. The independent test layout bypasses that defect. Physical print acceptance does not establish working authenticated persistence, household claiming, transfer, privacy or recovery. Those flows must be qualified before distribution.

Commission all four stations

  • Labels: correct the square-stock production exporter, verify identifier allocation and manifests, qualify the exact stock/tray, then test registration and scans on the actual printer. The saved independent 24-up file is a calibration aid, not an exporter repair.
  • Invitations: lock the assortment and paper dimensions; prove front/back alignment, color, dry time, crease quality and trim. Time 25 complete invitations, including assembly and packing, before extrapolating a 250-piece run.
  • Packaging: approve a fitted dieline or blank-mailer size. Check insert retention, adhesive cure, closure and the actual shipping package. For custom cutting, demonstrate the exact board, tooling, registration and finished boxes per hour on the quoted machine.
  • parts: qualify the same part/profile on both machines. Record dimensions, mass, time, rejects and bottle/insert fit. Retain the first accepted article; recheck after nozzle, material or profile changes.
  • Batch release: reconcile identifiers, inspect every finished unit for identity and visible defects, log sampled phone scans and product-specific fit/durability results, and count accepted/rework/void units before packing. No live identifier is silently duplicated to replace a reject.
  • Repeatability: require two consecutive representative batches to meet the documented criteria and measured schedule before advertising dependable throughput. Set an initial proposed 95% first-pass yield target, with every identifier discrepancy treated as a stop condition. Record actual performance; this is not a claimed result.

Work needed before fulfillment

TAGD owns the production-exporter and identifier lifecycle fixes. HomeHub owns its household/agent transfer and privacy workflow. Applied Sciences owns stock qualification, part design, assembly instructions and physical test evidence. Price only after materials, labor, machine time, waste, shipping, support and returns are accounted for. Unpassed product or software checks hold that product’s release; other qualified work can continue.