Introduction

What is VKRA?

VKRA is the monetization infrastructure for the Agentic Era. We provide a Retrieval-Augmented Generation (RAG) powered API that allows AI agents, LLM applications, and autonomous bots to discover and recommend relevant products contextually.

In a world where AI agents are becoming our primary interface for discovery and action, VKRA enables these agents to participate in commerce in a way that is helpful, high-intent, and privacy-preserving.

How it Works

VKRA operates on industry-standard semantic search patterns to deliver high-relevance recommendations.

1. Context Ingestion: You send the natural language context of your agent's conversation to our API.
2. Semantic Matching: We use vector similarity search to find the most relevant products from a massive database of verified retailers.
3. Contextual Response: We return a structured list of products, complete with metadata, relevance scores, and human-readable explanations.
4. Compliant Tracking: Every recommendation includes a unique identifier for privacy-preserving attribution tracking.

High-Level Architecture

graph TD
    A[AI Agent / App] -->|1. Context Query| B[VKRA API]
    B -->|2. Semantic Processing| C[Vector Knowledge Base]
    C -->|3. Relevant Products| B
    B -->|4. Contextual Results| A
    A -->|5. Interaction| D[Attribution Tracking]

Core Capabilities

Contextual Ad Search

The heart of VKRA is the /search endpoint. It handles natural language queries and returns high-intent product matches. It supports:

• Hybrid Filtering: Combine semantic search with hard constraints like price range, merchant, or brand.
• Progressive Streaming: Use Server-Sent Events (SSE) to deliver results in real-time as they are found.
• Batch Processing: Handle multiple intent-signals in a single efficient request.

Privacy & Compliance

• No PII: We never require or store Personally Identifiable Information.
• Intent-Based: Recommendations are based solely on the context of the interaction.
• Attribution First: Built-in support for major retail compliance requirements (Skimlinks, Amazon).

Developer Experience

• Simple REST API: Standard JSON patterns that work with any language.
• Extensive Metadata: Get everything you need to render beautiful product cards.
• Performance Focused: Low-latency responses designed for real-time agentic workflows.

Next Steps

• Quick Start: Make your first search request in minutes.
• Authentication: Learn how to secure your integration.
• Best Practices: Follow recommended patterns for privacy and performance.