How ScrapeOra works

Overview

ScrapeOra is a request-only, multi-tenant web-scraping platform for store owners — focused product sourcing without the dev work. You don't write scrapers; you browse a curated Scraper Store, run or schedule the ones you need, and sync the results straight into your storefront.

The headline flow: give a scraper a product listing URL, and ScrapeOra visits every product on it, extracts a normalized product record, optionally rewrites it with your own AI, and publishes it to your store — keeping it in sync on a schedule.

Scraper Store

The Scraper Store is the catalog of ready-made scrapers a Super Admin curates. You can browse and run any scraper available to you, but you can never author your own — scrapers are request-only. If your use case isn't covered, request a scraper from the Help Center and the team builds it.

The Overview page surfaces a Suggested Scrapers shortlist so you can get started in one click.

The scraper engine (two-phase)

Every listing scraper follows one fixed two-phase model. It is intentionally thorough, not shallow — that depth is what makes the output publish-ready.

• Phase 1 — Enumerate: open the listing/category URL you provide and collect every product's detail-page URL (following products, never keywords).
• Phase 2 — Visit: open each product page individually and extract the full record — image, product URL, title, price, discount, SKU, stock and every other available field.
• Output: a normalized product schema, identical across every source, so the same data can map to any supported platform.
• It runs as one job with shared concurrency, politeness, retries and a resume cursor; proxy-aware for sites that need it (the books.toscrape demo needs none).

Run flow

When you run a scraper, you first choose where the data is headed: just view & export, or push to a connected integration. That choice is remembered so the post-run preview pre-selects the right destination — the scrape itself stays generic and can target any platform later.

Integrations & supported platforms

Each platform has its own adapter and its own settings — a per-platform mapper that turns the normalized product into that platform's inventory format and calls its API correctly. Connect an integration by picking the platform, choosing the category to map into, entering your credentials (stored encrypted, per tenant), and optionally enabling AI.

Supported platforms: Shopify, WooCommerce, BigCommerce, Wix, Squarespace, Magento 2, PrestaShop, Ecwid, Medusa, plus generic Webhook and Google Sheets. Before publishing, products missing a required value (price, image, title) are filtered out — never published — and listed for you. Use the dry-run preview to see exactly what a push will create, update, skip or delist before it touches your store.

Schedules & inventory sync

A schedule is a scraper + listing URL, bound to one or more integrations, on an interval you choose (e.g. every 6 hours). On every tick it re-scrapes, maps, optionally AI-edits, and updates your store.

Sync is cross-run correct: each product is matched by a stable identity, so a re-scrape updates the same listing instead of creating a duplicate. Choose a sync mode per integration: additive (add + update, never delist), full sync (also delist products gone from the source), or price & stock only.

AI enrichment (optional, bring-your-own-key)

AI is optional and uses your own LLM key (OpenAI, Anthropic, Gemini or DeepSeek). When connecting an integration you can enable it and write, in plain language, what you want changed about each product. The rewrite is applied on push.

Guardrails: the model may restyle and clean copy but never fabricates specs or claims; raw price, SKU and images stay authoritative. Enrichment is cached, so re-pushing the same products doesn't re-spend your tokens. If AI is off, the raw scraped data is published unchanged.

Datasets (history & diff)

Every run's normalized products are stored and re-downloadable (JSON / CSV / platform formats). The Datasets view lets you compare any two runs of a scraper to see exactly what was added, removed, or changed (price, title, availability, image) — the same delta that powers scheduled sync.

Logs & notifications

Every push is logged in detail — what ran, what succeeded, what failed, and which products were skipped. Add a notification email per integration to get an email on both success and failure; the super-admin team is alerted on failures so they can step in.

Workspace, Accounts & Help Center

Workspace holds your AI providers. Accounts covers your team (seats + invites), API tokens, notifications and settings. The Help Center is your one stop for support tickets, your submissions, product limits, and requesting a new scraper.