Marketplace price parsing
A collector walks the listings you care about on a schedule and puts price, discount, stock and search position into one table. History is kept per platform.
- Buildfrom $1,400
- Timeline1–3 weeks
- Supportfrom $420/mo
Who this suits: Sellers and brands whose one item sits on three platforms at three prices, and who learn about a neighbour’s price cut from a drop in orders.
What is in the first version
- prices and discounts
- stock
- search position
- alerts
The figures above are the package this sits on: Parsing and reverse engineering. Anything past its edges the calculator adds before the work starts, not after.
How the path goes
- 01
Each run has a chosen region
Price and availability depend on region and warehouse, so each run fixes its region and records it beside the number.
- 02
The price breaks into parts
Price before discount, with discount, with the platform card and in a promotion are four fields, not one.
- 03
The change reaches the right person
A message arrives with the SKU, the old and new number, and what moved: seller price, platform discount or stock.
What decides whether this works or annoys people
A listing carries several prices
Repricing against a card price while the competitor number excludes it burns margin, so every record shows which numbers were compared.
Region changes price and stock
The same SKU costs differently by city and may have no stock, so region is a run parameter written into every row.
A vanished listing reads as no change
Silence reads as a price that held, so an empty answer is its own value and a vanished listing raises an alert.
What to measure once it is live
- share of listings collected first time
- time from a price change to the alert
- SKUs outside your price corridor
- repricings on the day of the change
Why this comes out faster
The shape of this one is known: the states, the edge cases and the things that usually go wrong have been decided before. Nothing here is a template, and the saving is not in your half of the work. It goes into your process, your content and the systems this has to talk to, which is the part nobody can have solved in advance.
A likely stack for this
Picked against the task when we scope it, not decided in advance. This is the shelf it usually comes off.
- CollectionPythonPlaywrightGoDocker
- Storage and historyPostgreSQLClickHouseRedisS3
- Schedule and alertsTemporaln8nTelegram Bot APIGrafana
Price it yourself, right here
Five steps, and you can see the number without leaving a contact. The estimate accounts for the kind of work, what you already have and what it will need inside.
These bills do not come from us
- proxies and captcha services in the regions you need
- paid tiers of the seller account and analytics
- AI model calls when listings match by name
- hosting, servers and space for the price history
Asked before the first call
- Why not through the marketplace API?
- A seller account gives up your data, not a competitor’s price or stock on the same listing. So your own comes via API, the rest off the public page, named per platform before the estimate.
- Can this put our seller account at risk?
- Collection happens from the buyer side, on public pages, and nobody signs into your seller account. If a figure exists only inside that account, we say so before the estimate: a different risk.
- Can it also set our prices on the platform?
- Reading and writing carry different risk, so the first version only reads. Writing a price back through the seller account is separate work, with a floor price and human confirmation on the first runs.