A data source is a URL that Skryx pulls product data from on a schedule. Once connected, Skryx auto-syncs the feed (every hour / 6 hours / daily) and keeps your search index in lockstep with your catalogue.
# Choosing a template
The Create-index wizard ships with pre-configured templates for the most common platforms. Each template knows the feed's field shape, so you don't have to map anything manually.
| Template | Format | Auth | Notes |
|---|---|---|---|
| Google Merchant | XML | Public | Standard g:* field map |
| Meta Catalog | XML | Public | Same g:* shape as Google |
| Shopify | JSON | Public or Custom App | /products.json?limit=250 |
| WooCommerce | JSON | Basic Auth | REST API /wp-json/wc/v3/products |
| Magento | JSON | Bearer token (custom header) | /rest/V1/products |
| PrestaShop | XML | Public | Google Shopping export module |
| Custom URL | Auto | Optional | Any XML, JSON, CSV or JSON-LD |
# Sync frequency
| Schedule | Use when |
|---|---|
| Hourly | Real-time inventory; tolerable extra fetches |
| Every 6 hours (default) | Most stores — best balance of freshness vs cost |
| Daily | Slow-changing catalogue |
| Manual | Push-driven (you trigger sync from the dashboard) |
# Testing a connection
The wizard's Test connection button fetches the URL once and shows:
- The detected format (XML / JSON / CSV / JSON-LD)
- How many products it parsed
- The first three sample products
- Any warnings (missing IDs, empty feed)
If the test fails, common causes:
- HTTP 404 / 403 — the URL is wrong, or it requires Basic Auth.
- Timeout — the feed takes more than 30 seconds to respond. Skryx aborts and retries on the real sync, but the test bails early to keep the wizard snappy.
- Auto-detect failed — the feed isn't valid XML, JSON or CSV. Use the Custom URL template if you've got a non-standard shape.
# After import
After the first sync brings in 50+ products, Skryx Auto-Pilot kicks in and configures language, synonyms, ranking and semantic search automatically. You'll be forwarded to the live progress page.
# Re-syncing manually
Open the index, find the data source row, and click Sync now. Skryx runs the sync immediately and records the result; the next scheduled run is unaffected.
# Failure handling
After three consecutive failed syncs, Skryx emails the workspace owner. After five, the data source auto-pauses so you don't get a notification storm. You can re-enable it from the index page once you've fixed the root cause.