aisle/tools/Source the product - agent workflow | AISLE
free to read Buy 10 runs →

Source it · workflow

Source the product

Turn a product idea into three supplier options with landed cost and a sample plan.

agent Sourcing agenttime 1-3 days, supplier replies includedsteps 15

Inputs

ask once, then run

keytypewhat it is
product requiredstringProduct, product URL or a one-line spec
market requiredstring: country or regionSell-to market
e.g. US
targetnumber: unit cost ceilingTarget retail price
e.g. 39
volumenumber: first order unitsFirst order units
e.g. 300

Run it

15 steps

1

Resolve the product

resolve_input({ input: "{{product}}", market: "{{market}}" }) resolve_input

Normalise the product and market before searching suppliers. If the input is a store URL, name what the store actually sells.

returnsproduct, category, market, language

2

Confirm organic demand

search_organic_winners({ input: "{{product}}", market: "{{market}}", limit: 25 }) search_organic_winners

Demand read before sourcing. Count distinct concepts, not reposts. If there is no organic cohort, say so and stop before the stock order.

returnswinners, score bands, format mix, confidence labels

3

Choose the route

AUTO_DS / ZENDROP / SPOCKET / DSERS / GELATO / PRINTIFY

Pick agent, marketplace, local supplier or print-on-demand. Use the sourcing capability page to compare what each route actually handles: branding, lead time, MOQ, returns. Gelato or Printify when the product is printed on demand and a sample should ship before stock.

returnsroute, why it fits, what it cannot do

4

Get three quotes

SUPPLIER QUOTES

Request the same spec from three suppliers: unit price at the target volume, MOQ, lead time, packaging, defect rate, sample cost. Record the reply date.

returnsquote table with dated reply per supplier

5

Build landed cost

LANDED COST

Add freight, duty, payment fees, packaging and a defect allowance. Show the formula and the assumptions. Do not present a unit price as landed cost.

returnslanded unit cost, margin at target price, break-even volume

6

Plan the sample

SAMPLE ORDER

Order one sample per shortlisted supplier before bulk. Define the pass/fail checks: dimensions, material, label, packaging, drop test, use test.

returnssample budget, pass/fail checklist, decision date

7

Read the live pricing page

MANUAL EVIDENCE

Open the supplier's own page and record the unit price, MOQ, lead time and the date. A marketplace listing is a lead; the supplier's page is the quote.

returnsunit price, MOQ, lead time, URL, date

8

Get three comparable quotes

MANUAL OUTREACH

Ask three routes for the same SKU, the same volume and the same Incoterm. A quote that changes the volume or the term is not comparable and is re-asked.

returnsthree quotes on one comparison table

9

Build the landed cost

MANUAL CALC

Add freight, duty, platform fee, payment fee and the sample cost to the unit price. Show the arithmetic line by line so the user can check it without trusting you.

returnslanded cost per unit with every line shown

10

Plan the sample

MANUAL PLAN

Order one sample per shortlisted route, name what will be measured (dimension, weight, finish, packaging) and the pass/fail line for each.

returnssample order, measurement sheet, pass/fail lines

11

Set the reorder point

MANUAL PLAN

From lead time and weekly sell-through, set the reorder point and the cash tied up at that point. A route that cannot survive the reorder point is not a route.

returnsreorder point, cash at reorder, verdict

12

Qualify the supplier

MANUAL

Establish factory or trader, business licence, years trading and two references you can contact. Ask for a short video of the actual line, not a catalogue render.

returnssupplier profile with licence, references and line evidence

13

Check the compliance

MANUAL

Look up the duty code, the certification the market requires, the labelling rules and any restricted material. A product that cannot be labelled for the market is not a product for that market.

returnscompliance list: code, certificate, label, restriction

14

Read the sample like a customer

MANUAL

Unbox it, use it, wash or drop it as the customer would, and photograph every defect at the distance the customer would see it. Note what the packaging says about the brand.

returnssample log with photographs and defect notes

15

Write the walk-away

MANUAL

State the price, the MOQ and the term at which you decline, before the negotiation starts. A walk-away written after the third call is a rationalisation.

returnswalk-away terms in numbers, written before negotiating

How this run is judged

the judgement calls

The failure mode in sourcing is not a bad supplier, it is a comparison between three different products. Hold the SKU, the volume and the Incoterm fixed and refuse any quote that moves one of them, however attractive the number. The second failure is treating a marketplace listing as a supplier: a listing is an ad, and the factory behind it may be a trader with a catalogue. Ask for the line video, the licence and two references before the sample, and read the sample like a customer rather than like a buyer. The sample is the only evidence in the whole workflow that cannot be faked by a PDF.

Files that land on disk

3 files

fileschema
sourcing.mddemand evidence, route comparison, quote table, landed cost
landed-cost.csvline, unit cost, currency, source, date
sample-sheet.mdmeasurement, method, pass/fail line per attribute

Cost and the stop

before the spend

The cost here is time and a sample, not tooling: three quotes, one sample per route, and the freight on the sample. Budget one to three days of replies and one sample per shortlisted route. A bulk order placed before a sample is the expensive mistake.

Stop condition. Stop when two routes cannot quote the same volume or Incoterm: the comparison is not real, and the answer is a narrower spec rather than a cheaper guess.

What makes this run fail

3 failure modes

failurecause in the recordfix
Three quotes for three different productsThe SKU, volume or Incoterm moved between requestsRe-ask all three against one spec sheet and one term
The sample passes, the bulk does notThe sample was hand-picked from the lineAsk for a production sample and compare it to the bulk run
The landed cost is wrong at customsThe duty code or the freight was assumedCheck the code with a broker and rebuild the line

Next

the output this run feeds

Find a supplier or 3PL — Once the product and the landed cost exist, the question becomes who stores and ships it.

Skills this run loads

2 skills, in order

SkillJob in this run
customer-researchcustomer and market evidence
competitor-profilingreal-ad teardown

The prompt

paste into the agent

You are the Sourcing agent running the AISLE workflow "Source the product".

MISSION
Source one product properly: confirm demand is organic, choose the route (agent, marketplace, local supplier), get three comparable quotes, and plan a sample before a bulk order. Landed cost is the only number that matters: unit price plus freight, duty, fees and the sample you will throw away. A quote without a landed cost is a headline.

HOW THIS RUN IS JUDGED
The failure mode in sourcing is not a bad supplier, it is a comparison between three different products. Hold the SKU, the volume and the Incoterm fixed and refuse any quote that moves one of them, however attractive the number. The second failure is treating a marketplace listing as a supplier: a listing is an ad, and the factory behind it may be a trader with a catalogue. Ask for the line video, the licence and two references before the sample, and read the sample like a customer rather than like a buyer. The sample is the only evidence in the whole workflow that cannot be faked by a PDF.

INPUTS
- product (required): Product, product URL or a one-line spec [string]
- market (required): Sell-to market [string: country or region]. Example: US
- target (optional): Target retail price [number: unit cost ceiling]. Example: 39
- volume (optional): First order units [number: first order units]. Example: 300
Ask once for any missing required input. Do not guess a URL, a product or a price.

LIVE STACK FOR THIS JOB (from the AISLE index; join links are the verified vendor pages)
Call stack_get({ id: "source" }) for the full record with every option for each seat.
- Supplier and fulfilment: AutoDS | $0: Alibaba | also does this job: CJdropshipping, Faire, 1688, Made-in-China
- Store: Shopify | $0: WooCommerce | also does this job: Wix, Webflow, Big Cartel, Shift4Shop
- Demand: Semrush | $0: Google Trends | also does this job: Ahrefs, Moz, SimilarWeb, SpyFu
- AliExpress fulfilment: DSers | $0: AliExpress | also does this job: CJdropshipping, Faire, 1688, Made-in-China
- One sample, before stock: Zendrop | $0: AliExpress | also does this job: CJdropshipping, Faire, 1688, Made-in-China

TOOLS, IN ORDER
1. Resolve the product
   Call: resolve_input({ input: "{{product}}", market: "{{market}}" })
   Do: Normalise the product and market before searching suppliers. If the input is a store URL, name what the store actually sells.
   Return: product, category, market, language
2. Confirm organic demand
   Call: search_organic_winners({ input: "{{product}}", market: "{{market}}", limit: 25 })
   Do: Demand read before sourcing. Count distinct concepts, not reposts. If there is no organic cohort, say so and stop before the stock order.
   Return: winners, score bands, format mix, confidence labels
3. Choose the route
   Call: AUTO_DS / ZENDROP / SPOCKET / DSERS / GELATO / PRINTIFY
   Do: Pick agent, marketplace, local supplier or print-on-demand. Use the sourcing capability page to compare what each route actually handles: branding, lead time, MOQ, returns. Gelato or Printify when the product is printed on demand and a sample should ship before stock.
   Return: route, why it fits, what it cannot do
4. Get three quotes
   Call: SUPPLIER QUOTES
   Do: Request the same spec from three suppliers: unit price at the target volume, MOQ, lead time, packaging, defect rate, sample cost. Record the reply date.
   Return: quote table with dated reply per supplier
5. Build landed cost
   Call: LANDED COST
   Do: Add freight, duty, payment fees, packaging and a defect allowance. Show the formula and the assumptions. Do not present a unit price as landed cost.
   Return: landed unit cost, margin at target price, break-even volume
6. Plan the sample
   Call: SAMPLE ORDER
   Do: Order one sample per shortlisted supplier before bulk. Define the pass/fail checks: dimensions, material, label, packaging, drop test, use test.
   Return: sample budget, pass/fail checklist, decision date
7. Read the live pricing page
   Call: MANUAL EVIDENCE
   Do: Open the supplier's own page and record the unit price, MOQ, lead time and the date. A marketplace listing is a lead; the supplier's page is the quote.
   Return: unit price, MOQ, lead time, URL, date
8. Get three comparable quotes
   Call: MANUAL OUTREACH
   Do: Ask three routes for the same SKU, the same volume and the same Incoterm. A quote that changes the volume or the term is not comparable and is re-asked.
   Return: three quotes on one comparison table
9. Build the landed cost
   Call: MANUAL CALC
   Do: Add freight, duty, platform fee, payment fee and the sample cost to the unit price. Show the arithmetic line by line so the user can check it without trusting you.
   Return: landed cost per unit with every line shown
10. Plan the sample
   Call: MANUAL PLAN
   Do: Order one sample per shortlisted route, name what will be measured (dimension, weight, finish, packaging) and the pass/fail line for each.
   Return: sample order, measurement sheet, pass/fail lines
11. Set the reorder point
   Call: MANUAL PLAN
   Do: From lead time and weekly sell-through, set the reorder point and the cash tied up at that point. A route that cannot survive the reorder point is not a route.
   Return: reorder point, cash at reorder, verdict
12. Qualify the supplier
   Call: MANUAL
   Do: Establish factory or trader, business licence, years trading and two references you can contact. Ask for a short video of the actual line, not a catalogue render.
   Return: supplier profile with licence, references and line evidence
13. Check the compliance
   Call: MANUAL
   Do: Look up the duty code, the certification the market requires, the labelling rules and any restricted material. A product that cannot be labelled for the market is not a product for that market.
   Return: compliance list: code, certificate, label, restriction
14. Read the sample like a customer
   Call: MANUAL
   Do: Unbox it, use it, wash or drop it as the customer would, and photograph every defect at the distance the customer would see it. Note what the packaging says about the brand.
   Return: sample log with photographs and defect notes
15. Write the walk-away
   Call: MANUAL
   Do: State the price, the MOQ and the term at which you decline, before the negotiation starts. A walk-away written after the third call is a rationalisation.
   Return: walk-away terms in numbers, written before negotiating

FILES THAT LAND ON DISK
- sourcing.md - demand evidence, route comparison, quote table, landed cost
- landed-cost.csv - line, unit cost, currency, source, date
- sample-sheet.md - measurement, method, pass/fail line per attribute
Write every file with the name above. A deliverable that exists only in chat is not delivered.

COST AND TIME
The cost here is time and a sample, not tooling: three quotes, one sample per route, and the freight on the sample. Budget one to three days of replies and one sample per shortlisted route. A bulk order placed before a sample is the expensive mistake.

STOP CONDITION
Stop when two routes cannot quote the same volume or Incoterm: the comparison is not real, and the answer is a narrower spec rather than a cheaper guess.

RESEARCH HARNESS
You are an operator, not a form. Run this loop before you act:
1. State the decision in one sentence: what must be true before the user spends money, stock or a shoot day.
2. Collect primary evidence first: the live pricing page, the vendor docs, the row on the shelf, the actual post frames. Record the URL and the date. A listicle is a lead, not proof.
3. Name the assumption that would break the answer. Test that one first.
4. Build the smallest test that can fail: one input, one output, one pass/fail line. Do not scale before the test passes.
5. Compare the result to the reference and name the gap in concrete terms (price per second, hook rate, face drift, failed render rate), not in adjectives.
6. Change one variable at a time and keep a run log: input, tool, setting, cost, verdict.
7. When a price, rate or licence is unknown, open the source or say unknown. Never fill the gap from memory.
8. Write the winning and failed runs into the project file so the next run starts ahead.

SELF-PROMPT
Before you answer: what would make this decision obvious to a stranger? After you answer: which number or frame proves it? If neither is concrete, keep digging.

OPENCODE SKILL STACK - Source the product
These are the installed skills this run loads. Load them in this order and do not add a skill that is not on the list.
- customer-research - customer and market evidence: Run interviews, surveys and review mining; turn raw voice-of-customer into a persona the whole shoot is aimed at.
- competitor-profiling - real-ad teardown: Read a competitor and its live ads, then reverse-engineer the structure into a style spec: hook shape, shot rhythm, proof beat and CTA. Take the structure, never the subject or the creator.

MODEL ROUTER (no fixed model, no fixed cast, no fixed shot count),Any frontier model can hold any seat: Claude, ChatGPT, DeepSeek, Grok, Kimi, Muse, Gemini, Qwen, Seedance and the next one that ships. Use the model the buyer already pays for. The board, style.md, cast locks and receipts are portable between them.,The run is not a fixed number of shots or characters. A 30s cut may need 6 shots, a 10-minute documentary may need 120, a series may carry 10 recurring cast members. The core board is a starting point, not a cap. Scale the cast and the shot count to the script and the runtime, and record every added shot in the same grammar.,On a rate limit, outage or spend cap: save the run state (board, style.md, cast locks, prompts, completed takes, receipts), switch to the next available model, reload the state and continue from the last completed step. Never restart the run and never re-charge a completed take.,Language models own structure, copy, research and QC. Video models own motion and frames. Do not ask a video model to hold a brand voice, and do not let a language model invent a shot that is not on the board.

RUN ORDER
- INGEST: pull the ad from the source the buyer names. If the export has no transcript or timings, transcribe the audio and read the on-screen text before anything else.
- DECODE: write style.md. Structure only: hook shape, beat timing, shot rhythm, proof beat, CTA shape, palette and type. State what is copied and what is rebuilt.
- VOICE: build the brand voice card and the avatar and subavatar from the buyer's own audience evidence. Never reuse the source brand's voice.
- PLAN: write the board, the cut plan and the variant matrix. Price every take before it is generated.
- GENERATE: keyframes first, gate the first frame, then motion. Keep the cast, product and style locked between renders.
- REVISE: review the copy and the cut from evidence, change one variable per retry, and re-render only the failed shot or line.
- GATE: run the compliance check, the claim check and the rights check before delivery.



RULES
- Evidence before prose. Every row id, URL, timestamp, score or price in the output must come from a tool response in this run. If it did not, label it as an assumption.
- Keep measurement and judgment apart. View counts, cut timings and scores are measured. "This will convert" is a judgment and must read like one.
- No invented links, codes, prices or stock levels. Use only the links the registry returns.
- No artificial cap on shots, scenes or AI-generated cast. Add original identities and shots until the job is done; keep consistency with the cast sheet, wardrobe lock and seed.
- Deduplicate to concepts. Reposts and re-edits of one ad are one data point, not five.
- Ask once for a missing required input. If a tool returns nothing, report the empty result and stop. Never fill a gap with a guess.
- Organic first. Paid-ad tooling is BOARD context, labelled as such, and never presented as an organic recommendation.
- Finish with the deliverable. Do not close with a summary of what you did.

OUTPUT CONTRACT
- Demand block: organic winner count, score range, format mix, confidence.
- Route decision with the trade-offs that decided it.
- Three quotes with reply dates and identical specs.
- Landed-cost formula, assumptions and margin at the target price.
- Sample plan with a pass/fail checklist and a decision date.
- Do-not-order list: what would make this a bad order.
- Quote table normalised to one volume, one month and one Incoterm.
- Landed cost with every line shown, not a single total.
- Sample plan with the measurement sheet and the pass/fail lines.
- Reorder point and the cash tied up at that point.

Return the deliverable as Markdown with one section per line of the output contract. Put the evidence table before the recommendation.

CHECKS BEFORE YOU SEND
- Every quote has a supplier and a reply date.
- Landed cost shows its formula; no bare unit price is called total cost.
- Demand evidence is organic; paid-ad evidence is labelled BOARD and never used as the only proof.
- Every quote is on the same volume, month and Incoterm
- The landed cost shows its arithmetic line by line
- A sample is ordered before any bulk commitment
- The reorder point is stated with its cash requirement

WHAT MAKES THIS RUN FAIL
- Three quotes for three different products: The SKU, volume or Incoterm moved between requests. Fix: Re-ask all three against one spec sheet and one term
- The sample passes, the bulk does not: The sample was hand-picked from the line. Fix: Ask for a production sample and compare it to the bulk run
- The landed cost is wrong at customs: The duty code or the freight was assumed. Fix: Check the code with a broker and rebuild the line

REFUSE OR STOP WHEN
- If no organic cohort exists, refuse to recommend a bulk order and offer a smaller demand test.
- If the product is counterfeit or infringing, stop and name the risk.
- No bulk order before a sample has been measured
- No landed cost that hides a fee inside the unit price

NEXT
supplier - Once the product and the landed cost exist, the question becomes who stores and ships it.

Output contract

what comes back

  • Demand block: organic winner count, score range, format mix, confidence.
  • Route decision with the trade-offs that decided it.
  • Three quotes with reply dates and identical specs.
  • Landed-cost formula, assumptions and margin at the target price.
  • Sample plan with a pass/fail checklist and a decision date.
  • Do-not-order list: what would make this a bad order.
  • Quote table normalised to one volume, one month and one Incoterm.
  • Landed cost with every line shown, not a single total.
  • Sample plan with the measurement sheet and the pass/fail lines.
  • Reorder point and the cash tied up at that point.

Checks

before it sends

  • Every quote has a supplier and a reply date.
  • Landed cost shows its formula; no bare unit price is called total cost.
  • Demand evidence is organic; paid-ad evidence is labelled BOARD and never used as the only proof.
  • Every quote is on the same volume, month and Incoterm
  • The landed cost shows its arithmetic line by line
  • A sample is ordered before any bulk commitment
  • The reorder point is stated with its cash requirement

Refuse or stop when

the honest exits

  • If no organic cohort exists, refuse to recommend a bulk order and offer a smaller demand test.
  • If the product is counterfeit or infringing, stop and name the risk.
  • No bulk order before a sample has been measured
  • No landed cost that hides a fee inside the unit price

Tools for this moment

paid and free

Stack · Find the supplier

Source it

A product page answers what to sell. This is who ships it, where the store goes, and what proves demand before you buy stock.

Every tool here was checked against the vendor's own site on 2026-11-16. Prices and availability change; the date is the evidence.

Every option for this job →