Product Matcher

Run a match

Start a match run for your group, see what it costs before it starts, and stop it if you need to. You start the run yourself. Before the run starts, the most it can cost is held in your wallet. If your wallet holds too little, the run does not start and nothing is charged.

In the portalVia APIVia MCP

1. Open your group in My Data, pick the Matching tab, then Match the data.
2. Check your lines between Your list and each Compared shop (see Set up a match). If you see Unsaved changes, press Save config first.
3. Open the 3-dots menu next to Run match and pick Run configuration. Match settings shows the Run summary (your list, the lines and the shops) and the Schedule.
4. Read the cost next to the button. It says “Up to” an amount “reserved” and what a run typically costs; the rest comes back after the run. Open How is this charged? to see each cost line.
5. Press Run match. The Run this match again? window opens with the cost. Under “What happens to your current matches”, pick what happens to your automatic matches and to your own decisions. The safe choice (“Keep and add new” and “Keep all”) is set each time. See Your current matches.
6. Press Run. The button shows “Queuing…” and then Running. The amount is now held in your wallet.
7. To end the run early, press Stop next to the button, then Stop run. To stop it and start a new run at once, press Run match again: the button in the window reads “Stop it and run”. See Stop it and run.
8. When the run ends, open Results. Rows are split into Matched, No match and Review. Your current matches stay visible until the new run finishes. If it fails or is stopped, nothing changes.

Note

The cost of the run is held in your wallet when it starts. When the run ends, you pay what it used and the rest comes back to your balance.

Note

If you press Run without changing anything on the canvas, the run uses your stored setup and your stored review levels as they are.

Your current matches

Each Run asks what happens to the matches you have now. There are two choices. The window starts at the safe choice every time: keep both.

• Old matches stay. Your current matches stay visible until the run finishes. If it fails or is stopped, nothing changes.
• What wins inside a run. For each product, a rejected pair is never matched. Then a pair you approved wins. Then, with “Keep and add new”, the old automatic match wins over a different new one. Only after that do the rules decide. Your choices never change scores, Floor, Max or lines.
• Decisions during a run. A decision you make while a run works is used by this run if you make it before the run starts writing results. Otherwise the next run uses it.
• Scheduled runs. If the group has a schedule, the window also shows “Use this automatic choice for scheduled runs”. Turn it on to use the same automatic choice for scheduled runs (not with “Start fresh”). Scheduled runs always keep your decisions. See Schedule and history.
• You changed your list. If you changed the main list, old matches stay until the run finishes. Decisions on products that are not in the new list stop applying, but they are kept. Over the API the answer has the warning MAIN_LIST_CHANGED. A shop you left out gives SOURCE_DROPPED: its matches leave when the run completes.
• Where a match came from. Each matched row has matchOrigin: VERDICT (you approved it), KEPT (kept from before) or ENGINE (made by this run). An older row without the field is ENGINE.
• Counts before you run. GET /api/scraper/group/{id}/match/config-status (Read) and scrapewise_get_matcher_config_status give automaticMatches (your current matches, not counting your approvals) and decisionCounts (approved and rejected, the decisions a “Start fresh” would set aside).

Stop it and run

While a run is waiting or running, the button in the run window reads “Stop it and run”. A warning above the choices says that the matching done so far in the current run is lost, and your delivered matches stay. The window says: “This replaces the current run.”

• API. Send PUT /api/scraper/group/{id}/match (Full) with "replaceRunning": true. The answer has the warning RUN_STOPPED: the run in progress was stopped and a new run was queued. Your delivered matches stay until the new run completes.
• MCP. scrapewise_update_matcher_job (Full) with replaceRunning set to true. Ask your assistant to do this only when you agree, because the work of the stopped run is lost.
• Without it. A Run while another run is busy can get 409 ENGINE_BUSY. Wait for the run to end, or send the Run again with replaceRunning.
• Cost. The new run holds money in your wallet like any run.

Costs

A run costs server time, rows and pictures. The estimate shows the cost before you start. In the portal, open How is this charged? next to Run match:

Search server M (cpx52, EUR [PRICE]/h): up to EUR [AMOUNT]

Rows: 50,000 × EUR 0.00015 = EUR 7.50

Pictures: 1,240 × EUR 1.50 per 1,000 = EUR 1.86

Pictures are counted only when photos are compared, see Photo check. The server price depends on the size of your group.

• Held, then settled. When the run starts, the most it can cost is held (“Up to … reserved”). When it ends you pay what it used; the rest comes back.
• Our fault, no charge. If the server fails to start, or the run stops because of a fault on our side, you pay nothing.
• You stop it. If you stop a run after its server is ready, the started hours are charged.
• Pictures in the estimate. The pictures line is based on your group’s last comparable run: rows × (pictures ÷ rows of that run) × 1.25, never more than one picture per row. With no earlier run, it counts one picture per row. The amount held when you press Run uses the same estimate, also with “Stop it and run”.
• History. After the run, the Cost column in History shows what you paid, for example “Server EUR 0.40 + rows EUR 0.10 + 1,200 pictures EUR 1.80”.

Errors you may see

{
  "code": "FLOW_RUN_SERVER_WALLET_LOW",
  "message": "Matching did not start: this run needs up to €1.93 and your available balance is €0.50. Top up €1.43 and run again. The previous result is kept.",
  "reserveMicros": 1932750, "availableMicros": 500000, "missingMicros": 1432750,
  "errors": [ { "code": "FLOW_RUN_SERVER_WALLET_LOW", "message": "...", "reserveMicros": 1932750, "availableMicros": 500000, "missingMicros": 1432750 } ]
}

The other money stop, 402 WALLET_INSUFFICIENT_BALANCE, is for scrapes and requests, not for match runs. See Wallet and billing.

Questions

Last updated 10 November 2026