For AI agents: this page is also available as Markdown at /docs/integrations.md. Every docs page is: add .md to its URL. llms.txt lists them all.
Integrations overview
WarpSearch connectors index a CMS or store catalog into a WarpSearch collection and serve the site's full-text search, filters and facets from it, with no search cluster to run. They use the same Elasticsearch-compatible endpoint as any other client.
| Platform | What you install | Status | Page |
|---|---|---|---|
| Umbraco 18 | NuGet package WarpSearch.Umbraco.Search | Available | Umbraco |
| WooCommerce | WordPress plugin (GitHub) | Available | WooCommerce |
| Magento 2 / Adobe Commerce | Composer package warpsearch/module-search (GitHub) | Available | Magento |
| Shopify | WarpSearch app and theme block | Coming soon | Shopify |
| BigCommerce | WarpSearch single-click app | Coming soon | BigCommerce |
The Umbraco, WooCommerce and Magento connectors are open source and installed on your site. Shopify and BigCommerce are managed apps that WarpSearch hosts; they are awaiting marketplace review.
#Operation and deployment ownership
| Platform | Indexing and search | Who deploys and operates it |
|---|---|---|
| Umbraco | The Umbraco Search provider indexes content, media and members, including culture variants and access filters. Umbraco APIs and backoffice search use WarpSearch. All site indexes share one collection. | The site's developer installs the NuGet package, configures its collection/key, deploys the Umbraco site and starts the initial rebuild. WarpSearch operates the search service. |
| WooCommerce | Product and stock hooks write parent product documents with variation SKUs/attributes. Action Scheduler or WP-Cron retries failures. REST, shortcode/block and optional native product-search replacement fetch IDs and render current WooCommerce data. Rebuilds sweep old generations after refresh. | The merchant or agency installs/configures the plugin on WordPress, runs a rebuild and maintains scheduled jobs. No separate connector server. |
| Magento / Adobe Commerce | A document per product/store view, with parent/child fields and guest pricing. Product, stock, category and mass-update events feed a database queue. A quick-search proxy hydrates Magento products. This is not a replacement for Magento's full search adapter/layered navigation. | The Magento operator installs the module, runs setup/DI compilation, configures the key/collection and operates the consumer/cron. Periodic full reindex covers changes that bypass hooks. |
| Shopify | Bulk export builds the catalog; authenticated webhooks re-fetch changed products and inventory. The managed service polls bulk completion, retries durable jobs and reconciles daily. A theme block searches through Shopify's signed app proxy. | WarpSearch deploys the managed backend and Shopify configuration/theme extension. Merchants authorize the app, confirm their WarpSearch organization and add the theme block. |
| BigCommerce | Paginated catalog import, product/SKU/category webhooks and daily reconciliation update the catalog. Script Manager loads a Stencil search widget; its requests use the managed backend's search proxy. | WarpSearch registers and deploys the app/backend. Merchants authorize it and confirm their organization; setup registers hooks and installs the script. |
Once live, Console > Integrations provisions a dedicated collection and server-held key for each connected Shopify or BigCommerce store, shows synchronization status, and offers resync/disconnect. Merchants need no separate hosting.
#Shared setup
Site-installed packages and standalone development apps need a collection and a key. Create both once in the console. Managed Shopify/BigCommerce onboarding provisions these for you:
- Create a collection, for example
mysite_prod. Leave the mapping empty: the connectors name their fields so WarpSearch types them automatically. - Create an API key with the writer role on that collection.
Use one collection per site and environment. A connector can keep several indexes in one collection, so the free plan is enough to start.
Connectors talk to https://api.warpsearch.io, send the key as Authorization: Bearer ws_..., write with POST /compat/es/{collection}/_bulk and search with POST /compat/es/{collection}/_search. They retry rate-limited and failed requests with backoff and honor Retry-After.
Keep the key out of source control and out of the browser. The store connectors search through a server-side route that holds the key, so shoppers never see it.
#Plans
Connectors run on the free tier or on Starter, the flat monthly plan for live sites. The summary is on WarpSearch for e-commerce and CMS sites; the parts that affect a connector:
| Free | Starter | |
|---|---|---|
| Collections | 1 | 5, so staging and production, or several sites, stay separate |
| Freshness | New and changed documents are searchable within a few minutes | Within seconds, including price and stock changes |
| Traffic | 5 searches a second, 1 at a time; the connector shows a "try again" state when the limit is hit | 50 a second, 10 at a time |
| Warmth | The index loads on the first search after a quiet period | Kept warm |
| Relevance | Defaults | Synonyms, pinned results and boosts from the console |
| Analytics | 7 days | 90 days, with export |
| Widget | Shows "Search by WarpSearch" | No badge |
| Idle | A collection with no traffic for 30 days is deleted, with notice | Kept |
Freshness is set by the platform's commit schedule, not by the connector, so every connector gets the same behavior on each plan. Connectors write as soon as something changes on either plan.
Your plan also sets the documents per write request, hits per search, facets per search and requests per second. Keep connector batch sizes, page sizes and facet counts within them. GET https://api.warpsearch.io/v1/usage shows your limits. See limits and errors and the cost guide.
#Not covered
The connectors use keyword search with filters and facets. They do not use vector or hybrid search yet. For that, call the REST API directly.