Site parsing is hands-free: Elbuz crawls a supplier's or competitor's site and puts names, prices, availability, photos, and specifications into the source price list, then processes it as usual.
When you need this:
- No price list, but the products are on the supplier's site — the wizard or AI picks the rules.
- The site has a ready-made parser — Rozetka, PROM, Hotline selectors are already written.
- Prices update nightly — server mode runs on a schedule.
- Wrong data came in — the parsing log shows where the chain broke off.
Where to start: connect the site with the source wizard, “Site parser” tile.
What the site parser can do
- Connect a site as a source — with a ready-made module from the list or with the universal parser by the site address.
- Set up parsing rules — show where on the page the name, price, and photo are; the wizard or AI picks the selectors.
- Use a ready-made selector — and XPath examples for the name, price, SKU, photo, and availability.
- Run collection manually or on a schedule — in your browser or in the background on the Elbuz server.
- Install the Elbuz extension for Chrome — it opens the pages when collection runs without the server.
- Troubleshoot a failure — use the parsing log to find the broken link in the chain, a CAPTCHA, or a site refusal.
How a parser differs from a regular price list
For the rest of the app, a parser source is just another price list: it has products, prices, markup, category matching, and a run log. The only differences are where the data comes from and what you set up.
| File, link, email, FTP | Site parser | |
|---|---|---|
| Where the data comes from | one file from the supplier | site pages the app crawls by itself |
| What you set up | file column addresses | page parsing rules — s |
| The “Check” step in the wizard | yes: column matching | skipped; the parser settings window opens instead |
| Running | “Import” reads the file | “Import” opens the “Import data from the site” window, and crawling starts right away |
| Logs | the source run log | plus the “Parsing log” and a live collection feed |
How crawling works: four operations
The app never takes “the whole site” at once. It uses a queue: one page, one . Each parsed page creates next-stage tasks: home page → sections → product pages → specifications. So the settings are four selector sets, one per operation, and an empty result almost always means one link broke off, not that “the parser doesn't work”.
| Operation | What the app does | What you set | What appears after it |
|---|---|---|---|
| List of product category links | opens one page (the home page or a catalog page) and collects section addresses from it | a selector for section links | a “List of product links” task for each section found |
| List of product links | opens a section — a — and collects product page addresses and next page addresses | a product block selector, a pagination selector, and the name and price fields if needed | a “Product card” task for each product and tasks for the next pages of the section |
| Product card | opens the product page and fills in the product fields, photos, videos, and files | field selectors: name, price, SKU, description, photo, and others | a product in the source price list; if needed, tasks for the specifications page and for option products |
| Attributes | parses the specifications table — on the same product page or on a separate page | a specifications block selector, a row selector, name and value selectors | product specifications |
The list of operation types has a fifth one, “Photo”, but no rules are created for it. Photos are collected by the image field in “Product card” or in “List of product links”.
Scenarios
First collection from a supplier's website
Add the source with the wizard: the “Site parser” tile, the site address in the “Universal parser” block (how it works). With auto-setup, the first run collects no more than 300 products and stops by itself. Check the result in the source price list, add selectors for the missing fields (parsing rules), and start a full collection (running).
Collection used to work, but this time the wrong data came in
Open the “Parsing log” before the next run: the response code and page size show whether the site refused, or the page came in but the selectors no longer parse it. Next, test the rule and pick the selectors again (order of checks).
The same site — for yourself and for a competitor
Two sources on the same ready-made module share one set of settings: a change in one changes the other too. Add the second source for this site with the universal parser — it has its own rule set (about ready-made parsers).
FAQ
Do I need to keep the window open while collection runs?
In browser mode, yes: the extension opens the pages in your browser, and closing the window stops collection. In server mode, you can close the window. More: two collection modes.
Do I have to install the Chrome extension for the parser to work?
No. You only need it when your browser opens the pages. With the “Use server to load data” module, the Elbuz server downloads the pages, and native sources never need the extension. How to install it: Elbuz extension for Chrome.
The parser for a supplier's website doesn't import anything. Where can I see what's going on?
In the “Parsing log”: source settings window → “Additional” tab → “Parsing log”. Code 200 and a normal size mean the problem is in the parsing rules; 403, 429, or a very small size mean the site didn't return the page. More: parsing log.
Can the parser sign in to a website?
No, sign-in forms can't be set in the settings window. The interface only gives you cookies and HTTP headers in the “Additional” accordion at the bottom of the “Additional” tab (collection settings).
Can I collect articles or news from a website instead of products?
Yes. The “Additional” accordion on the “Additional” tab has the “Type of items to load” switch: “Product”, “Article”, “News”, “Recipe”. The default is “Product” (collection settings).
Can the system send a Telegram message when the import from a supplier's website is finished?
Only for the whole price list queue: in “My profile & security”, turn on “Products imported from all price lists” — you get an email, Telegram, and push. Or use the “Notification” action in Flows after “Import price lists”. Server-side may still be running at that moment.
Where to find it and who has access
None of the parser windows is in the menu or has a desktop shortcut. All three open from one place — the price list list or the card of the source itself.
- Menu
- Directories → Processing price lists → List of price lists
- From the source card
- “Source” section → “Site data import settings” button
- From the source list
- a row of the “Site parser” type → “Setup” button in the right panel header
- From the connection wizard
- “From where” step → “Site parser” tile → “Next”
- Access right
- the “Price lists” window; for the parsing log, also “Parsing log” or “Website parsing log”
- Prepare
- a contact and the site address; to collect without the server — the Elbuz extension for Chrome
- Module
- “Use server to load data” — for background collection and scheduling
Rights are set for the window as a whole: you can't give access “to the link list only, not to the settings”. The parser settings window and the collection window open for anyone who can open the source card; the parsing log checks its own right separately.
1
List of price lists- List of price lists — the only door to all site parser windows.
The button and the window it opens have the same name — “Site data import settings”; the access rights list uses it too. Below, it's simply the settings window.
In the source list, a parser row has a globe icon and the “Site parser” type. If you don't enter your own name at the “Who” step, the universal is named after the site domain; a ready-made module keeps the name the wizard gave it.
1
Site parser source row2
Setup3
Import- Source row — the globe icon marks a site parser source, unlike file-based sources.
- Setup (gear) — appears only for site parser sources and opens the site data import settings.
- Import — for a parser it opens the collection window and starts the crawl; not pressed.
Related topics
- If you're comparing the parser with other ways to get a price list — Connecting a supplier source: file, link, email, , .
- If you need to understand what happens to the collected products next — Importing a supplier price list: stop rules and operations after crawling.
- If you need the numbers of each run — Price list log.
- If you need to know exactly which fields the parser takes — What exactly the parser takes from a product page.
