graphewayBETA

Your site opens its own door to AI agents

v0.2.2Public Beta

Agents Stop
Scraping
Your Site

Grapheway turns your website into a live, typed knowledge graph that AI agents discover, traverse, search, and act on. One package, zero dependencies, no paid crawlers. Your site opens its own door.

$npm install grapheway
See a Probe RunView GitHub Repo
Interactive Knowledge Graph31 Nodes · 44 Connections
LAYER:
ACTIVE AGENTExplore the 3D knowledge graph by dragging with your cursor, or select any node to inspect details

One Probe, a Whole Agent Surface

Real commands and real output. Every snippet below is the exact format the project writes to disk.

terminal, grapheway probe
$ bunx grapheway probe https://expressjs.com

  Probed Express.js (https://expressjs.com)
    pages:    25
    headings: 521
    edges:    1927  (1361 extracted, 566 inferred)
    api:      12 endpoints (https://expressjs.com/openapi.json)

  Serving "Express.js" as an agent surface:
    discovery    http://localhost:4321/.well-known/agent
    graph        http://localhost:4321/graph/v1
    events       http://localhost:4321/graph/v1/events (realtime SSE)
    MCP          http://localhost:4321/mcp

# That legacy site is now an MCP server. Point Claude Desktop
# or Cursor at localhost:4321/mcp and its docs become tools.

Surface

Every capability between install and a working setup. Search by keyword, filter by category.

find$
category
GRAPHTYPED NODES / EDGES

A Live Knowledge Graph of Your Site

Pages and sections become typed nodes, links become edges with provenance. Agents call graph_node, graph_neighbors, graph_search, and graph_path instead of scraping HTML.

MCPSTREAMABLE HTTP

Open MCP, No API Key

The /mcp endpoint is a spec-aligned Model Context Protocol server. Claude Desktop, Cursor, VS Code, and any MCP client get your graph tools and actions as native tools, plus pages as markdown resources. No key, no account.

PROBEANY LEGACY SITE

Probe Any Website Into a Graph

Point grapheway probe at any URL, no site involvement needed. It extracts the site's own knowledge, title, navigation, headings, links, and OpenAPI endpoints, then serves the same agent surface locally or exports it as JSON.

DISCOVERYA2A-STYLE CARD

Agents Find You at /.well-known/agent

An A2A-style agent card tells any agent exactly what your site exposes, how, and where. Discovery is a URL, not a partnership negotiation.

ACTIONSHTTP + MCP

Let Agents Do Things

Your basic actions, like check status or query, become callable tools over HTTP and MCP. Same implementation, both surfaces.

COMPATLLMS.TXT / SITEMAP

Legacy Files, Decoupled

grapheway/compat optionally serves llms.txt, agents.txt, robots.txt, and sitemap.xml at runtime, for the agents that still probe them. Fully decoupled: mount it or do not.

RUNTIMENODE / EXPRESS / HONO

Runs Anywhere, Zero Deps

The core is framework agnostic and dependency free. It runs in Node, Bun, and Deno, works with Express, Hono, Next.js, plain node:http, or static hosts. The handler interface is four fields in, four fields out.

One Probe of Express.js

Real numbers from running grapheway probe against expressjs.com with default settings. Your numbers will differ per site.

SurfaceWhat agents gotProvenanceFreshness
grapheway probeLEADING25 pages, 1,927 edgesextracted + inferredlive, re-crawl on demand
Scraping HTML25 downloads per agentmarkup, untypedstale on next change
llms.txt snapshotone flat filehand maintainedstatic until edited
Paid crawler APIsame pagesthird partyon their schedule
●Probe Verified

Under the Hood

Six subpath exports, about 6,000 lines of TypeScript, zero dependencies. Where the code lives.

MODULE 1

core

1,336 lines. The graph model: typed nodes, typed edges, properties, versioning. Graph operations, live patching with structural validation, diff between snapshots, the discovery manifest, and Schema.org JSON-LD serialization.

MODULE 2

web

2,154 lines. The drop-in agent endpoint: createGrapheway, 14 routes across discovery, graph, viewer, agent API, and MCP. Framework adapters for node:http, Express, and Hono. Live SSE events, patchGraph, and GEO head injection.

MODULE 3

probe

985 lines. The agent-side crawler: fetches any site same-origin only, respects robots.txt, extracts nav, headings, links, and OpenAPI specs into tagged edges, then serves or exports the result.

MODULE 4

compat

540 lines. The optional legacy module: llms.txt, agents.txt, agents.json, robots.txt, and sitemap.xml served at runtime. Fully decoupled from the core.

MODULE 5

agent

385 lines. The typed GraphewayClient for agents: discovery, graph summary, node, edges, search, path, native traverse, manifest, pages as markdown, and action calls.

MODULE 6

cli

661 lines. The grapheway command: probe, gateway, serve, mcp-config, audit, and generate. mcp-config prints the exact client snippet to paste.

CONTRIBUTEAll development happens publicly on GitHub. The white paper in docs/whitepaper.md explains the architecture, and examples/simple-site is a working site you can run in one command. Contributions follow the guide in CONTRIBUTING.md.
Read the white paper: the problem, the insight, the architecture
docs/whitepaper.md
DOCS
Run the example site: one bun command, full agent surface
examples/simple-site
EXAMPLE
The agent skill: SKILL.md ships in the package
packages/grapheway/SKILL.md
AGENT
Changelog follows Keep a Changelog with public history
CHANGELOG.md
CHANGELOG
SUPPORT

Keep the Gateway Open

Grapheway is GPL-3.0 and built in the open, with a white paper and a public changelog. If your agents stopped scraping today, star the repo or pass it to another web team.