Search Page & Embed Code

Search Page & Embed Code

The hosted search page is a free, fully-featured search UI at search.opensolr.com/your-index-name — use it as-is or embed it into your own website with a single line of code. No server setup, no coding required.

What Your Search Page Looks Like

Below is a visual overview of the hosted search page. It has three main areas: the search bar at the top, a facet filter sidebar on the left, and the results list on the right.

Your Search Page URL

Every index you create on Opensolr automatically gets its own public search page. The URL follows a simple pattern:

Search Page Features

Your hosted search page is packed with features out of the box. No configuration needed — they all work automatically based on your index data.

Content Tabs

The search page organizes results into three tabs at the top. Users can switch between them to find exactly the type of content they need.

Sorting Options

By default, results are sorted by relevance — the best matches appear first. Users can also sort by date to see the newest content first. For ecommerce indexes with a price field, price sorting is available too. For a deep dive into the underlying Solr parameters, see Search Parameters Explained.

Fresh Results Bias

Next to the sort controls there is a Fresh toggle. It is off by default. Turn it on and every result's score is multiplied by a recency curve on the document's creation_date — full weight for something published today, about half after a year — so newer documents climb the ranking.

The important part is what it does not do: it never filters. The hit count is identical either way, nothing older becomes unreachable, and a document with no date at all simply keeps its place rather than being pushed to the bottom. That is what separates it from a date filter, which removes results outright.

The choice belongs to the person searching, not to whoever set the index up: the same index answers “the most relevant thing” and “the most relevant recent thing” depending on which the visitor asks for. It travels in the URL as fresh_bias=1, so a fresh-biased search can be bookmarked and shared, and it survives paging, facet clicks and a new query typed into the box.

It works on every search shape — keyword-only, vector-only and the fused hybrid ranking — because the boost wraps the final score rather than one half of it. On the API and in the client packages the same behaviour is available as a fresh_bias parameter.

AI Search Toggle

On vector-capable indexes, an AI toggle sits next to Fresh. Its starting position mirrors the index's Search Mode from Index Settings: on for any hybrid mode, off when the index is set to Lexical Only.

The visitor can flip it either way, and their choice overrides the index setting for their own searches only — the index configuration never changes. Switched on, the search runs as full hybrid (keyword + semantic vector matching, union mode) even on a Lexical Only index. Switched off, the search runs as pure keyword matching even on a hybrid index — useful when someone wants exact, deterministic keyword results with no semantic neighbours.

The choice travels in the URL as ai=yes or ai=no, so it can be bookmarked and shared, and it survives paging, tab switches, spellcheck suggestions, facet clicks and new queries. In the embed widget the same control is available as the data-ai attribute.

Search Operators

The ? button beside the search box opens a short reference to the operators a visitor can type. They work identically with the AI toggle on and off.

They compose in one query: +laptop +"13 inch" -refurbished returns only 13-inch laptops and never a refurbished one.

Anything with + or - in front of it — a word or a phrase — becomes a filter over the whole result set, so an exclusion holds even when the semantic half of a hybrid search would otherwise have surfaced that document. An unprefixed phrase influences ranking on the keyword side instead; write +"…" when you need it enforced. Ordinary hyphenated words such as e-mail and covid-19 are searched for literally. The reasoning is on the Hybrid Search page.

Search by Image

To the left of the search box there is a small image button. Click it, pick a photo, and the search runs on what is in the picture. On a phone the same button opens the camera, so a visitor can photograph something and search for it on the spot.

Photo in, results out. There is nothing to switch on, nothing to add to your index and nothing to re-index — and everything you already set up keeps working exactly as it did: ranking, facets, filters, freshness, query elevation and your saved Search Tuning.

The words themselves stay out of the way. The search box shows a small thumbnail of the uploaded picture instead, so the visitor can see which image the results belong to; hovering enlarges it and clicking opens it full size. The query is still in the URL, so an image search can be bookmarked and shared like any other, and typing in the box replaces it with a normal text search.

Labels, receipts and documents are read, not just recognised. When the picture is mostly text — a shelf tag in a store, a spec label on a box, a receipt — Opensolr reads it and searches with the largest print on it plus any article or model number it finds, which is how a photo of a door's shelf tag lands on that exact door. Because a query like 6048297 G1-88 is something to match rather than something to understand, that search runs with the AI toggle off and the toggle switches back to the visitor's own setting on their next typed query. Under the results the page says what it read, so the visitor always knows why they got what they got.

Accepted: JPEG, PNG, WEBP, GIF, BMP, TIFF, HEIC and AVIF, up to 20 MB. The uploaded picture is never stored — it is described, then discarded. The same capability is available to your own applications through the image_to_text API endpoint.

What it looks like. Nine real searches, on live indexes — click any of them to open it full size.

Embedding Search in Your Website

You can embed the full Opensolr search experience directly into any page on your website. It takes one line of code. The embedded search widget looks and behaves exactly like the hosted page, but lives inside your site.

The Embed Code

Copy and paste this single line into any HTML page where you want the search widget to appear:

Embed Code Attributes

Each attribute controls a different aspect of the embedded search widget. Here is a full reference:

Theme Options: Light vs Dark

The embed code supports two visual themes. Pick the one that best matches your website's look and feel.

Pagination Styles

Choose how users navigate through large result sets. Both styles work with all themes and layouts.

Tips for Embedding

Related FAQ Articles

Related Documentation