> ## Documentation Index
> Fetch the complete documentation index at: https://guide.beenos-solutions.com/llms.txt
> Use this file to discover all available pages before exploring further.

# Search Your Beenos Solutions Knowledge Base Content

> Query your Beenos Solutions knowledge base using natural language, keyword, or hybrid search — directly from the dashboard or via the REST API.

Beenos Solutions includes a powerful built-in search engine that lets you query your knowledge base directly from the dashboard — no agent required. Use it to verify that the right content is indexed, to test how your agents will respond to specific questions, or simply to find a passage buried inside a large document. Search supports natural language queries, exact keyword matching, and a hybrid mode that blends both for the best coverage.

## Using the Search UI

To open the search interface, navigate to your knowledge base and click the **Search** tab at the top of the page. A search bar appears along with a results panel that updates as you type.

Refine your results using the filter controls beneath the search bar:

* **Source** — limit results to one or more specific sources. Select sources by name from the dropdown. This is useful when you know the answer lives in a particular document and want to reduce noise.
* **Date added** — filter results to sources added before, after, or between specific dates. Use this to exclude older content that may no longer be accurate.
* **Content type** — narrow results to a specific source type: **Document**, **URL**, or **Text**. This helps when you have mixed sources and want to focus on a particular input format.

All filters are additive — results must match every active filter simultaneously.

## Search Types

Beenos Solutions offers three search modes. You can switch between them using the mode selector next to the search bar.

**Semantic search** interprets your query as a concept rather than a literal string. It uses vector embeddings to find passages that are *about* what you're asking, even if they don't contain your exact words. Use semantic search when you're asking a question in natural language, such as *"What happens if a customer misses a payment?"*

**Keyword search** performs exact token matching against the indexed text. It behaves like a traditional full-text search engine and is most reliable when you're looking for specific terms, product names, error codes, or identifiers that must appear verbatim in the results.

**Hybrid search** combines semantic and keyword signals and ranks results by a weighted relevance score that considers both dimensions. This is the default mode and works well for the widest range of query types — it catches both conceptually related passages and exact-match content in a single pass.

## Reading Search Results

Each result card in the results panel shows:

* **Relevance score** — a numerical score between 0.00 and 1.00 indicating how well the chunk matched your query. Scores above 0.75 generally indicate a strong match. Scores below 0.40 suggest the result is loosely related and may not be useful to your agent.
* **Source attribution** — the name of the source the chunk came from, along with its type icon. Click the source name to open the source detail panel.
* **Snippet** — a short excerpt of the indexed text showing the matched passage. Keywords from your query are highlighted in bold within the snippet. For semantic matches, the most relevant sentence in the chunk is surfaced even if your query terms don't appear literally.

If a query returns no results, try broadening your phrasing, switching to semantic mode, or checking that the relevant source has an **Indexed** status badge.

## Using Search via the API

You can query your knowledge base programmatically using the Beenos Solutions REST API. This lets you integrate search into your own applications, run automated tests against your index, or build custom retrieval workflows outside of the dashboard.

Send a `POST` request to the `/v1/query` endpoint with your query string and optional parameters:

```bash theme={null}
curl -X POST https://api.beenossolutions.com/v1/query \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"query": "What is our refund policy?", "limit": 5}'
```

The `limit` parameter controls the maximum number of result chunks returned (default: `5`, maximum: `20`). The response includes each chunk's text, relevance score, and source metadata. You can also pass a `knowledge_base_id` field to scope the query to a specific knowledge base when your account contains more than one.

Replace `YOUR_API_KEY` with a key generated from **Settings → API Keys** in your dashboard.

<Tip>
  Write queries the way a user would phrase a question, not as a list of keywords. Queries like *"How do I cancel my subscription?"* return more relevant semantic results than *"cancel subscription"*. For hybrid and semantic modes especially, full questions with context outperform short keyword fragments.
</Tip>