Yavy
Yavy · Coding · Other
Yavy is an AI knowledge platform that turns any website, documentation set, or private content source into an AI-searchable knowledge base. It sits between your content and large language models, indexing your pages so AI assistants pull real answers from what you actually published instead of guessing. Think of it as a managed RAG alternative, minus the pipeline you'd normally have to build yourself. You can reach that index through the Yavy CLI, a portable Skills package, or an MCP server connected to tools like Claude, Cursor, and VS Code.
About Yavy
What Is Yavy
Yavy is a knowledge platform built for teams who want their AI assistants to answer from source material rather than from memory. You point it at a content source, and it crawls, chunks, and embeds the pages using semantic search with vector embeddings so that any connected assistant can retrieve the exact passage a question needs. When an assistant asks a question, it gets back clean JSON grounded in your own docs. No more pasted-in PDFs or half-remembered answers.
The pitch is narrow on purpose. Yavy doesn't try to be a chatbot or a writing app. It's the retrieval layer underneath those tools, so the answers your AI gives trace back to your actual content.
The catch is that it's a paid product. Plans start at $9 per month, and there's no free tier sitting under them, so you'll want a real use case before signing up. It also leans toward developer workflows, since the CLI and MCP setup assume you're comfortable in a terminal. Not a dealbreaker for engineers. Worth knowing for everyone else.
Getting Started
- Create a Yavy account and verify your email to reach the dashboard.
- Add a content source: a website URL, GitHub repository, Notion workspace, or Confluence space.
- Let Yavy discover and index your pages automatically, then review the project in the dashboard.
- Install the CLI with npm install -g @yavydev/cli and run yavy init to authenticate.
- Search your docs from the terminal with yavy search "your query", or connect an AI tool through the project's MCP server URL.
Product Information
A quick look at Yavy's pricing, supported platforms, and performance.
Best for
The users, tasks, and scenarios where this tool fits best.
Users
- Developers
- Technical writers and docs teams
- Small product teams
Tasks
- Grounding an AI assistant in product docs
- Searching internal documentation
- Generating portable Skills packages
Scenarios
- Onboarding new engineers who ask "how does this work"
- Customer support where reps need fast, accurate lookups
- Legal and compliance reference
Key features
Semantic Search with Vector Embeddings
Yavy embeds your content so search matches on meaning, not keywords. If a user asks about "resetting a password" and your docs say "recover your account," the assistant still finds the right page because the retrieval layer compares concepts rather than exact strings, which is exactly the gap keyword search leaves open. That matters most when your terminology doesn't match what people actually type.
MCP Server for Real-Time Access
Every project gets a unique MCP server URL. Connect it to Claude Desktop, Cursor, Windsurf, VS Code, or ChatGPT and the tool searches your content live. The standard MCP protocol means it works with any assistant that supports it, not just one vendor's tools. Simple as that.
CLI and Portable Skills
The Yavy CLI lets you search docs from the terminal and generate portable Skills packages. Drop those Skills into a project and the AI loads them as local knowledge, no server connection required. Claude Code users can run /yavy:init to set the whole thing up automatically, which handles install, authentication, and skill generation in one command. Handy if you'd rather not wire it up by hand.
Automatic Discovery and Indexing
Add a URL and Yavy finds your pages on its own, using sitemap detection and recursive crawling across whatever the source exposes. It also supports GitHub, Notion, and Confluence as sources, so you're not limited to public websites. Paste in a repo link and it pulls the content straight from there.
Context-Aware Chunking
Yavy splits content by semantic headers rather than fixed paragraph blocks, so each chunk keeps the heading structure that told a reader what the section was about in the first place. The result is that each retrieved chunk carries a coherent idea instead of a fragment cut mid-thought, which cuts down on the noise your assistant has to filter before it can answer a real question. So why does chunking matter this much? Because a bad split hands the model half a sentence and it fills the gap with guesswork.
Always-Fresh Re-Indexing
Automatic re-indexing keeps the knowledge base current as your docs change. On the Pro plan this runs hourly, so an updated page shows up in answers without a manual rebuild. No stale context. That's the difference between an assistant that cites yesterday's policy and one that cites the version you published this morning.
Pros and cons
Pros
- Answers stay grounded in your actual content, which is the whole point for teams tired of hallucinations.
- Three ways to connect, CLI, Skills, and MCP, so it fits different workflows instead of forcing one.
- Supports private content sources, including authenticated projects for internal material.
- Broad AI tool coverage, with guides for Claude, Cursor, Windsurf, VS Code, ChatGPT, and more.
Cons
- No free plan means you can't test the full workflow without paying first.
- The setup leans technical, since the CLI and MCP path assumes terminal familiarity.
- The product is early, still in its public beta cycle, so expect features to shift.
Frequently asked questions
It turns a website or docs set into a searchable knowledge base for AI assistants. Instead of an assistant guessing, it retrieves real passages from your content and hands them to the model as context, so the answer a user sees is drawn from something you actually wrote rather than a plausible-sounding guess.
Related content
Explore related tools, skills, and articles for Yavy.
Yavy Alternatives
Base Code
Base44 (Wix) · Coding · ProductivityBase Code is a shared cloud development environment from Base44. Teammates describe changes to an existing GitHub codebase in chat, preview them live and send them back as pull requests.
Google EmbeddingGemma 2
Google DeepMind · Writing · Image · CodingGoogle EmbeddingGemma 2 is a 740M-parameter open model that turns text, code, images, audio, and video frames into vectors in one 768-dimension space, running on phones and laptops.
GitGlow
GitGlow · Coding · ProductivityGitGlow is a local-first desktop Git client that lets independent developers review their own code and their AI agents' changes in one place before shipping.