Skip to content

About the Open LLMO Research Initiative

The Open LLMO Research Initiative is an independent research initiative on AI retrieval, citation, and grounding for the open web. It publishes specifications, benchmarks, and open-source tooling as its primary outputs.

Research AI-native retrieval, grounding visibility, and LLM-oriented information architecture, and publish reproducible metrics and specifications.

AreaScope
AI Citation AnalysisThe conditions under which LLMs cite content, and citation frequency
Grounding VisibilityMethods for visualizing what AI bases its responses on
LLM Retrieval OptimizationDocument optimization for the LLM retrieval layer
AI-native DocumentationResearch on document formats that LLMs handle well
Agent-oriented Information ArchitectureInformation structures that AI agents can operate on

The LLMO / AEO / GEO space is expanding rapidly, but three foundational pieces are missing:

  • No reproducible measurement — there is no public tool that plays the role of Lighthouse or PageSpeed Insights for AI discoverability
  • No shared vocabulary or scope — each vendor publishes its own definitions and the field has fragmented
  • Little open experimental data — commercial SEO tools dominate, and the research layer is thin

This Initiative is built to fill those three gaps. The goal is to play the role Lighthouse plays for SEO: publish the methodology, ship the tooling, and let the community build on top.

PrincipleMeaning
Reproducibility firstEvery metric ships with a calculation formula and an OSS checker
Draft over StandardSpecifications are published as “Draft / Experimental / Proposal v0.1” so they remain revisable
Open Source firstTools under OSS licenses, data under CC BY, specifications under MIT
Solo-honestSolo operation is stated explicitly rather than dressed up as a consortium

Ken Imoto. Author of multiple books on LLMO and harness engineering, published on Zenn and Amazon Kindle. Founder and CEO of Propel-Lab Inc. Responsible for the implementation and operation of multiple in-house frameworks and llmoframework.com.

Main publications:

The Initiative matures in phases. Each phase is the prerequisite for the next.

PhaseScopeStatus
Phase 0Research framing, Mission publication, first Experiment LogIn progress
Phase 1Reproducibility — OSS CLI (llmo-checker), Score v0.1 Draft, dataset publicationPlanned
Phase 2Community — contributors, external references, feedback channelsPlanned
Phase 3Standardization — formal specifications, Compatible certification badge, Working Group formationPlanned

Standardization comes last. Without mature OSS, benchmarks, and implementations to back them up, neither certification nor specifications can earn trust.

MethodLink
Issues / bug reportsgithub.com/kenimo49/llmo-guide/issues
Pull Requestsgithub.com/kenimo49/llmo-guide

This site and all draft specifications are published under the MIT License.