Engawa
Enables websites to expose a first-class interface for AI agents by providing read-only MCP tools like search_site and generating llms.txt discovery files.
README
Engawa
The open toolkit for agent-native websites.
Bring your agent.
Engawa helps websites expose a first-class interface for AI agents alongside the human-facing site. Visitors can bring the agent they already trust instead of using another embedded chatbot.
Why Engawa
Traditional websites optimize for humans browsing HTML. Agents scraping that HTML get noisy, incomplete context. Engawa sits between your site and the agent ecosystem—like an engawa, the transitional space between inside and outside in Japanese architecture.
WEBSITE
│
┌────────────┴────────────┐
│ │
Human interface Agent interface
│ │
HTML/UI Engawa
│
┌────────┼────────┐
│ │ │
llms.txt MCP content
What it does
- Validates generic site configuration (
EngawaConfig) - Normalizes content through pluggable adapters
- Generates llms.txt v2 discovery files
- Exposes MCP resources and safe read-only tools (e.g.
search_site) - Documents an implementation profile for agent-native websites
Current status
Early v0.1 foundation. Not production-hardened. API may change before 1.0.
Quick example
See examples/minimal-site for Example Studio—a fictional site with llms.txt, markdown pages, and an MCP endpoint.
pnpm install
pnpm build
pnpm --filter minimal-site start
Requires Node.js 24 LTS (see CONTRIBUTING.md).
Packages
| Package | Purpose |
|---|---|
@thierry-gilgen-ict/engawa-core |
Config, resources, adapters |
@thierry-gilgen-ict/engawa-discovery |
llms.txt generation, link metadata |
@thierry-gilgen-ict/engawa-mcp |
MCP server adapter (spec 2026-07-28) |
Planned (not in v0.1): React UI, Next.js adapters, CLI, analytics.
Security defaults
Public Engawa endpoints are read-only by default in v0.1. No unauthenticated mutation, no secret access, bounded search input and results. See docs/security-model.md.
Licensing
Roadmap
See docs/roadmap.md.
Contributing
See CONTRIBUTING.md.
License
MIT for software. CC BY 4.0 for docs. Copyright Thierry Gilgen ICT, 2026.
Recommended Servers
playwright-mcp
A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.
Magic Component Platform (MCP)
An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.
Audiense Insights MCP Server
Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.
VeyraX MCP
Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.
graphlit-mcp-server
The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.
Kagi MCP Server
An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.
E2B
Using MCP to run code via e2b.
Neon Database
MCP server for interacting with Neon Management API and databases
Exa Search
A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.
Qdrant Server
This repository is an example of how to create a MCP server for Qdrant, a vector search engine.