> ## Documentation Index
> Fetch the complete documentation index at: https://docs.valyu.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Design Philosophy
> Understanding Valyu's architecture and intended use cases
Valyu is built on a simple idea: AI systems need access to authoritative knowledge to give reliable answers. This page explains how we've designed Valyu and what it's best at.
## Design Principles
### 1. Built for AI
Valyu is designed for AI agents and LLMs, not adapted from a traditional search engine.
**What this means:**
* Semantic understanding, not keyword matching
* Structured JSON responses ready for LLMs
* Embedding-powered retrieval that understands context
* RAG-optimised formatting with citations and metadata
**Why it matters:**
Traditional search APIs return links for humans to click. Valyu returns structured data that AI can use directly—reducing hallucinations and improving accuracy.
### 2. One API, Many Sources
Valyu unifies multiple authoritative data sources into a single search interface.
**What you can search:**
* **Real-time web** - Current events and trending topics
* **Academic papers** - Research-backed insights
* **Books and literature** - Domain knowledge
* **Financial data** - Real-time market information
* **Proprietary datasets** - Specialised industry data
**Why we did this:**
AI agents need diverse, authoritative sources. Rather than integrating multiple APIs yourself, you get one comprehensive retrieval layer that spans the knowledge spectrum.
### 3. Transparent Pricing
Every search has predictable, granular pricing with limits you control.
**How it works:**
* **Pay-per-use CPM pricing** - No subscriptions
* **Max price controls** - Set spending limits per query
* **Relevance thresholds** - Filter out low-quality results
* **Source-specific pricing** - Premium sources cost more, but you choose when to use them
**Our philosophy:**
You should have complete control over costs without sacrificing quality.
## What Valyu Is Good At
### Ideal Use Cases
**Retrieval-Augmented Generation (RAG)**
* Grounding LLM responses with authoritative sources
* Reducing hallucinations
* Providing citations
**AI Research Assistants**
* Literature reviews
* Cross-referencing sources
* Finding recent research
**Knowledge-Heavy Chatbots**
* Educational assistants
* Professional services (legal, medical, financial)
**Real-Time Information**
* News and current events
* Market data
* Time-sensitive decisions
**Specialised Search**
* Academic research
* Financial analysis
* Technical documentation
See our [data coverage](/concepts/data-coverage) page for what's available and
the [Prompting Guide](/search/prompting) for best practices.
## What Valyu Isn't For
### Less Suitable Use Cases
**Direct User Search**
* Users querying the API directly (not through an LLM)
**Social Media Monitoring**
* Twitter/X trends
* Sentiment analysis
* Viral content
**E-commerce**
* Shopping comparisons
* Product reviews
* Inventory checks
**Local Search**
* Restaurant recommendations
* Local services
* Location-based queries
**Creative Content**
* Image generation prompts
* Creative writing
* Entertainment recommendations
## Build on Top of Valyu
Rather than trying to solve every search use case, Valyu provides precise, comprehensive search that you can compose into your own workflows.
## What's Next
Valyu is evolving based on developer feedback:
* **Better precision** - Especially for fuzzy queries and exact document matching
* **Agentic search framework** - Build your own search workflows
* **Multimodal search** - Images and documents
* **Custom sources** - Private data connectivity
* **Advanced filtering** - More granular controls
**Feedback?** Reach out at [contact@valyu.ai](mailto:contact@valyu.ai)
## Get Started
Get searching in minutes
Write better queries