# Doofinder MCP for AI Agents AI Agent Connect

> Doofinder gives your AI agent direct access to your e-commerce search engine. You can query products, audit search analytics, and check your index health without jumping between tabs. It's the fastest way to see how customers are finding your items while making real-time adjustments to your search logic and filters.

## Overview
- **Category:** ecommerce
- **Price:** Free
- **Endpoint:** https://edge.vinkius.com/vk_preview_imo70mJQ10SSxIq8oe3Qgc4G2DqYyLls9BR0IrTl/ai-agent-connect
- **Tags:** site-search, search-analytics, predictive-search, conversion-optimization, e-commerce-tools

## Description

When you're trying to figure out why a specific brand isn't showing up in search results, you usually have to hunt through a dashboard to see how your filters are behaving and whether your sort orders are actually doing what you intended. This Connector changes that by letting you just ask your agent for the data. You can query the search engine history or pull live stats on query velocity without ever leaving your chat window. It takes the guesswork out of search optimization. When you're trying to figure out why 'summer shoes' aren't hitting the right pages, you can check the click-through rates instantly to spot the bottleneck. It turns a tedious audit into a quick conversation. By connecting Doofinder to your workspace via the Vinkius catalog, you're basically giving your agent a seat in the control room. You can check your product graph, verify your tenant indexes, and grab predictive nodes to see what customers are typing before they even hit enter. You move from 'I think our search is working' to 'I know exactly how our search is performing' without the manual overhead.

## Tools

### search_filtered
Applies brand, color, or price filters to narrow down SKU results. Use this to see how your facet logic behaves in real time.

### get_search_engines
Pulls the history of your search engine gateways for validation. It helps you catch routing errors before they affect customers.

### get_indices
Shows the active arrays and auth status of your search indices. Use this to verify your search engine is healthy.

### search_custom
Extracts specific product data based on custom validation rules. This helps you see exactly what matches your unique search criteria.

### get_items
Pulls raw catalog data from internal arrays to check product structures. This is great for auditing your product graph.

### get_stats
Retrieves click-through rates and query velocity from hold parsing. It gives you a clear picture of search performance.

### search_keyword
Finds specific records within the headless Doofinder platform. Use it to see exactly what a specific keyword returns.

### search_pagination
Shows cloud logging data for specific vault limits. This helps you understand how your search results are being paginated.

### search_sorted
Generates product lists with specific sort directions like price or relevance. Test different sorting logic without manual changes.

### suggest
Exports the rules for predictive search nodes and billing data. See what customers see as they start typing.

## Prompt Examples

**Prompt:** 
```
Why aren't my 'blue dresses' showing up first?
```

**Response:** 
```
I checked the search results for 'blue dresses' using your filters. It looks like the 'relevance' sort is currently prioritizing price over color. You might want to adjust the sort direction in your Doofinder settings.
```

**Prompt:** 
```
Give me the search stats for the last week.
```

**Response:** 
```
Here are your search stats from the last 7 days:

| Metric | Value |
| :--- | :--- |
| **Total Queries** | 14,500 |
| **Avg. CTR** | 7.2% |
| **Query Velocity** | +3% |

**Top Queries:**
* summer sandals
* linen shirts
* swimwear
```

**Prompt:** 
```
What are the top suggestions for 'iph'?
```

**Response:** 
```
For the partial query 'iph', your search engine is currently suggesting:

* iphone 15
* iphone case
* iphone screen protector
* iphone charger
```

## Capabilities

### Query products with custom keywords
Find specific items using custom search logic and bounded records.

### Apply multi-facet filters
Narrow down results by brand, color, price, or other account properties.

### Retrieve search analytics
Get real-time click-through rates and query velocity numbers.

### Preview predictive suggestions
See what customers are typing via your active predictive search nodes.

### Audit index health
Inspect internal arrays to verify your product graph and data structures.

### Validate search gateways
Check your gateway history to ensure search requests are routing correctly.

## Use Cases

### Fixing a 'no results' page
A marketer notices a brand isn't showing up. They ask the agent to use search_filtered to check if the brand facet is broken.

### Auditing new arrivals
A product owner wants to see if new items are indexed. They ask the agent to use get_indices and get_items to verify the product graph.

### Optimizing sort orders
A manager wants to prioritize high-margin items. They use search_sorted to test different price and relevance directions.

### Analyzing search trends
A developer wants to see what people are searching for. They use get_stats to pull query velocity and click counts.

## Benefits

- Stop guessing about search performance by using get_stats to see real-time CTR and query velocity.
- Fix broken navigation faster with search_filtered to see exactly how your facet filters are behaving.
- Audit your entire product graph in seconds using get_items to ensure your catalog data is accurate.
- Improve conversion rates by checking predictive nodes with suggest to see what customers are actually typing.
- Validate your search engine setup quickly with get_search_engines to catch gateway errors before they affect sales.
- Test complex search queries without manual clicks using search_keyword to see exactly what your customers see.

## How It Works

The bottom line is you get a direct line of communication between your AI agent and your e-commerce search data.

1. Subscribe to the Doofinder MCP on Vinkius
2. Enter your Search Zone, HashID, and API Key into your client
3. Ask your agent to pull stats or test search filters in plain English

## Frequently Asked Questions

**Can Doofinder MCP help me see why products aren't appearing?**
Yes, it lets your agent check your search indices and filter logic to find out exactly why certain items are being skipped or hidden.

**How does Doofinder MCP handle search analytics?**
It pulls real-time data like click-through rates and query velocity so you can see how customers are interacting with your search results.

**Can I use Doofinder MCP to test my search filters?**
You can ask your agent to apply specific brand, color, or price filters to see the results instantly without having to do it manually in the dashboard.

**Does Doofinder MCP work with my existing Doofinder account?**
Yes, you just need your Search Zone, HashID, and API Key to connect your account to your AI client.

**Can Doofinder MCP help with predictive search?**
It can pull your active predictive nodes so you can see what suggestions your customers are seeing as they start typing.

**Is Doofinder MCP good for auditing my product graph?**
It allows you to inspect deep internal arrays to verify the data structures that define how your products are mapped and displayed.

**Can my agent perform filtered searches using specific product attributes?**
Yes. Use the 'search_filtered' tool. You can specify the 'filter_name' (e.g., 'brand', 'color') and the 'filter_value'. The agent will analyze the global bounds and return results restricted strictly to those custom limits.

**How do I monitor the click-through rate (CTR) of my search engine via chat?**
Use the 'get_stats' tool. Your agent will pull raw status configurations and capture exact CTR, total clicks, and query velocity numbers for your search engine, helping you track performance without manual dashboard checks.

**Can I check which search engines and indices are active in my account?**
Absolutely. The 'get_search_engines' and 'get_indices' tools allow your agent to dump all isolated tenant indexes mapping explicit hash strings, giving you a full view of your Doofinder infrastructure.