> For the complete documentation index, see [llms.txt](https://vitrina-blogerov.gitbook.io/bloggers-showcase/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://vitrina-blogerov.gitbook.io/bloggers-showcase/product-docs/product-docs-en/admin-panel/search-agent.md).

# Search agent

The search agent section manages AI search on the storefront. This AI helps the user describe a task in plain language and get suitable bloggers from the catalog.

<div data-with-frame="true"><figure><img src="https://162603775-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2FuwCSpQuu1p5soxSClCET%2Fuploads%2FLFzAzsmVdGP2pZyeEtFQ%2Fai-search-section.png?alt=media&amp;token=78f50f02-6173-412d-8751-4ae919d7eaa5" alt=""><figcaption></figcaption></figure></div>

{% hint style="success" %}
AI search does not replace the database. It works on top of existing cards, tags, prices, and descriptions.
{% endhint %}

### What can be configured

Available in the section:

* AI provider;
* API key;
* model identifier;
* system prompt;
* reset the prompt to the default version;
* saving changes.

### How a user query passes through AI

```mermaid
flowchart TD
    A["The user enters a query on the storefront"] --> B["AI search"]
    B --> C["Prompt + available search tools"]
    C --> D["Search by cards, tags, price, and followers"]
    D --> E["List of suitable blogger IDs"]
    E --> F["The storefront displays the found cards"]
```

### Provider and model

The platform supports OpenAI-compatible AI providers. In the admin panel, you can choose a provider and specify the model that the search agent will use.

{% hint style="danger" %}
For production use, it is strongly recommended to use `Groq`. The other providers are available as compatible options, but they have not gone through a full testing cycle in this release.
{% endhint %}

{% tabs %}
{% tab title="Provider" %}
Defines the external AI service the platform connects to. Recommended value: `groq`.
{% endtab %}

{% tab title="API key" %}
Gives the platform access to the selected AI service. The key is secret.
{% endtab %}

{% tab title="Model" %}
Specifies the specific AI model for processing requests. Recommended base model: `llama-3.3-70b-versatile`.
{% endtab %}
{% endtabs %}

### System prompt

The prompt explains to the agent how to search for bloggers, which tools to use, and how to return the result.

It can be adapted to the market: communication style, brand priorities, local specifics, typical client tasks.

{% hint style="warning" %}
Don't turn the prompt into arbitrary marketing copy. It should help the AI search the database correctly, not invent bloggers or promise what isn't in the catalog.
{% endhint %}

### Reset prompt

The reset button restores the default prompt. This is useful if search quality has worsened after experiments.

### What to check after configuration

After changing AI settings, it is worth checking several real queries:

* a query by topic;
* a query by budget;
* a query by platform;
* a query by audience;
* a complex query with multiple conditions.

### Connection to hidden tags

The search agent receives information about categories and tags from the database. Therefore, hidden tags can be used as internal features for AI selection: they do not clutter the user interface, but help the agent understand the card more accurately.

{% hint style="info" %}
For example, the user should not see the tag `avoid-finance`, but AI can take it into account as an internal constraint when selecting bloggers for financial advertising.
{% endhint %}


---

# Agent Instructions
This documentation is published with GitBook. GitBook is the documentation platform designed so that both humans and AI agents can read, navigate, and reason over technical content effectively. Learn more at gitbook.com.

## Querying This Documentation
If you need additional information that is not directly available in this page, you can query the documentation by asking a question.

Perform an HTTP GET request on the following URL with the `ask` and `goal` query parameters:

```
GET https://vitrina-blogerov.gitbook.io/bloggers-showcase/product-docs/product-docs-en/admin-panel/search-agent.md?ask=<question>&goal=<user_goal>
```

`ask` is the immediate question: it should be specific, self-contained, and written in natural language.
`goal` is what the user is ultimately trying to achieve, the reason they need the answer. Sharing it helps GitBook give you a better, more relevant answer. A goal is most helpful when it describes the outcome the user wants rather than restating the question. For example, with `ask=how do I create an API token`, a goal like `automate deployments from our CI pipeline` lets GitBook tailor the answer to that use case.

The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.