The open standard for agent navigation

Search finds pages. OpenNav makes them usable. Add the open navigation layer agents can discover before they waste tokens parsing visual HTML.

Get started View docs

llms.txt llms-full.txt /.well-known/opennav.json

The problem

The web was built for people. Agents need a map.

Today an agent opens a page, reads navigation chrome, scripts, layout, repeated headers, sidebars, and hidden presentation noise just to find the one answer it came for.

Raw HTML

Agents load the whole visual page.
Tokens burn on layout and repeated chrome.
Every redesign can break extraction.
Results vary from site to site.

OpenNav

Agents discover the readable interface first.
Markdown mirrors remove presentation noise.
Compatibility metadata tells agents what exists.
Navigation becomes predictable across sites.

Why OpenNav

The open standard for agent navigation

OpenNav sits between discovery and execution: after search finds a page, OpenNav gives agents the files, metadata, and readable routes they need to use it reliably.

Agent-ready in one build step

Add OpenNav to your deploy flow and publish the same site with an agent-readable layer beside the human-facing pages.

Open source by default

No sales cycle and no black box. Developers can inspect the output, wire it into frameworks, and help define the standard.

Less token waste

Agents read focused Markdown and metadata instead of paying to parse full HTML pages built for human eyes.

No custom API required

Most websites will never build a dedicated agent API. OpenNav gives them a universal compatibility layer they can ship now.

The position

Search finds pages. OpenNav makes them usable.

AI search, browsers, RAG systems, and APIs all solve different parts of the web access problem. OpenNav owns the missing layer: making the live web compatible with agents before they start reading.

Start with the docs