OKF Generator, Free Tool
Build your Open Knowledge
Format bundle in minutes
Google's Open Knowledge Format (OKF) gives AI agents a structured, machine-readable map of your company. Fill in the form below and Faro generates a complete /okf/ directory you can drop into your site in minutes.
Company information
What is OKF?
The emerging standard for machine-readable company knowledge
Google Cloud published the Open Knowledge Format (OKF) v0.1 in June 2026, a vendor-neutral spec for packaging business knowledge as directories of markdown files that AI agents can navigate. Think of it as your company's Wikipedia, structured so that ChatGPT, Claude, Gemini, and automated procurement agents can reliably read and reason over it.
Every file in an OKF bundle represents a single concept: a product, a process, a FAQ, or a pricing tier. with YAML frontmatter that gives agents context about what type of knowledge it is, when it was last updated, and where to find the live version. Concepts link to each other, forming a knowledge graph rather than isolated pages.
OKF is not a ranking factor, it does not directly affect search results. But it does determine how well AI agents can represent you in answers, comparisons, and procurement decisions. A company with an OKF bundle gives agents structured, authoritative facts to cite. A company without one forces agents to guess from unstructured web pages, or skip the company entirely.
The agent knowledge stack
Where OKF fits in the 5-layer model
AI agent readiness isn't one thing, it's a stack of capabilities. Most companies are stuck at layers 1 and 2. OKF gets you to layer 3, where structured knowledge starts replacing guesswork.
Layer 1: Discovery
Let agents find your content
Examples: sitemap.xml, robots.txt, llms.txt
Layer 2: Entity Understanding
Tell agents who you are
Examples: Schema.org Organization, sameAs links, author profiles
Layer 3: Agent Context
YOU ARE HEREStructured knowledge agents can reason over
Examples: OKF concept bundles, this tool
Layer 4: Actionability
Let agents take actions on your behalf
Examples: MCP endpoints, APIs, booking integrations
Layer 5: Governance
Make knowledge trustworthy and auditable
Examples: Provenance metadata, freshness signals, review status
Why this matters now
AI agents are the new procurement layer
Agents skip unreadable companies
When a procurement agent evaluates vendors, it reads structured data, not marketing copy. If your knowledge isn't structured, you're not in the consideration set.
OKF is early, advantage window is open
Most companies don't have an OKF bundle. The businesses that build one now will have a significant head start when agent-driven discovery becomes the norm.
Complements your existing setup
OKF doesn't replace Schema.org, llms.txt, or your sitemap. It extends them, adding the structured concept layer that agents need to reason accurately about your business.
Five years ago the question was “why do I need SEO?” Today everyone knows the answer. The same shift is underway for AI agent readiness. The question businesses will be asking in the next few years is “why isn't ChatGPT recommending us?”. The answer will often trace back to structured knowledge not being present or not being machine-readable.
OKF readiness checklist
What Faro checks in your AI readiness scan
Faro's AI readiness scan includes a dedicated OKF check. Here's what we look for:
FAQ
Common questions about OKF
Is OKF an official Google standard I have to use?
OKF v0.1 was published by Google Cloud in June 2026. It's a vendor-neutral open spec, not a requirement. You don't have to adopt it, but companies that do will be more legible to AI agents that use it for knowledge retrieval.
Will having an OKF bundle improve my SEO?
OKF is not a search ranking factor. Its primary benefit is AI agent discoverability, helping LLMs and automated systems accurately represent your business in answers, comparisons, and procurement decisions. Think of it as structured data for agents, not for search crawlers.
How often should I update my OKF bundle?
Whenever your company knowledge changes: new products, pricing updates, changed contact info. The log.md file tracks these changes. Faro's AI readiness scan flags concept timestamps older than 90 days as a potential staleness issue.
Do I need an OKF bundle if I already have an llms.txt?
They serve different purposes. llms.txt tells agents which pages to prioritize. OKF gives agents structured, typed concepts they can reason over directly. llms.txt ideally references your OKF index, so agents can go from 'here are my important pages' to 'here is my structured knowledge graph.'
How does this generator differ from building OKF manually?
The generator produces valid, spec-compliant OKF files in under 2 minutes. Building manually requires understanding the YAML frontmatter spec, getting the directory structure right, and writing each concept file from scratch. The output is identical, the generator just handles the boilerplate.
See your full AI readiness score
Your OKF bundle is one piece of the picture. Run a full Faro scan to see how your site scores across all 7 AI readiness categories, and get a fix for every failing check.
Run your AI readiness scan →