Yavy

What is Yavy?

Yavy is an AI knowledge platform that turns any website into an AI-searchable knowledge base. It acts as a bridge between your site and large language models (LLMs), enabling researchable, context-rich answers drawn directly from your content. When a user asks a question, the agent queries your content via the MCP protocol to fetch the exact context needed, reducing AI hallucinations and keeping answers anchored to your own material.

In addition to semantic search, Yavy embeds your content and returns structured data in clean JSON, optimized for LLM consumption. This enables robust, context-aware responses that scale across your knowledge bases.

Key Features

• AI-Ready Indexing: Automatic discovery and indexing of pages across your site and content sources.
• Live Sync: Real-time updates with milliseconds-wide propagation when content changes.
• Semantic Search: Vector embeddings that capture meaning, not just keywords, enabling concept-based results.
• Structured Context: Clean JSON context delivered to your AI agents for reliable responses.
• MCP Protocol: Plug & Play integration with any MCP client; standard endpoint per project.
• Public & Private Sources: Public sources work instantly; private sources use OAuth 2.1 for secure access.
• 100% Clean Data: We parse and normalize content to provide a consistent knowledge base.
• Zero Maintenance: Automatic indexing and content intelligence keep your AI up-to-date.
• Content Intelligence: Monitor competitor sites, blogs, docs for updates; real-time alerts on changes.
• Simple Pricing & Onboarding: Clear plans with a 30-day money-back guarantee.
• Flexible URL Discovery & Crawling: Automatic URL discovery, with sitemap-based crawling and recursive URL discovery.
• Platform Integrations: Ready-to-use MCP servers; once configured, you can connect to Claude, OpenAI, Windsurf, etc.
• Security & Access: OAuth 2.1 authentication, enterprise-grade security for private content.

How to Use Yavy

From URL to AI-ready in four simple steps:

1. Input: Provide your website or content URL.
2. Parse: We crawl and extract meaningful content from every page.
3. Index: Content is transformed into a searchable knowledge base.
4. Serve: Your AI assistants query content via the MCP standard.

Configuration is plug-and-play. Connect Yavy to your favorite MCP client in seconds. Just add the server configuration to your settings file:

• Open your agent configuration.
• Paste the server definition.
• Restart your agent.

Public sources work instantly. Private sources use OAuth 2.1 for secure access. The platform provides a standard MCP server endpoint for each project; copy your unique server URL from the dashboard and add it to your AI tool MCP configuration.

Use Cases

• Customer support automation: Build a bot that answers questions directly from your product docs, manuals, and knowledge base.
• Internal knowledge base: Keep teams aligned with up-to-date policies, API docs, and developer guides.
• Compliance and policy monitoring: Track changes across competitor sites, blogs, and documentation, with real-time alerts.
• Developer tools & API docs: Create chat-assisted lookup for API references, SDKs, and guides.
• Content-driven e-commerce or SaaS help centers: Improve self-service by linking product pages to user questions.

FAQ

• How does the search actually work? We use AI-powered semantic search with vector embeddings. This means your AI assistant finds relevant content based on meaning, not just keywords—even when the exact terms differ from the query.
• How quickly can I get started? Minutes. Add your website URL, and we immediately start discovering and indexing your pages. Most sites are fully searchable within a few minutes of signing up.
• How often is my content re-indexed? We automatically check your content for changes. Pro plans get daily re-indexing to keep your AI assistants current. You can also trigger a manual sync anytime from the dashboard.
• What types of websites are supported? Any website with a sitemap.xml — documentation sites, blogs, knowledge bases, and more. We support multiple discovery methods including sitemap-based crawling and recursive URL discovery. Works with any static site generator or CMS.
• Can I filter what gets indexed? Absolutely. You can configure allow/deny patterns in your dashboard to control exactly which pages are crawled and which are ignored.
• How do I connect this to my AI assistant? We provide a standard MCP server endpoint for each project. Just copy your unique server URL from the dashboard and add it to your AI tool MCP configuration—works with Claude, ChatGPT, and any MCP-compatible client.
• What counts toward usage? Usage is measured by the number of indexed pages and searches performed; you can monitor limits in the dashboard.