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.
Mission
Section titled “Mission”Research AI-native retrieval, grounding visibility, and LLM-oriented information architecture, and publish reproducible metrics and specifications.
Research Areas
Section titled “Research Areas”| Area | Scope |
|---|---|
| AI Citation Analysis | The conditions under which LLMs cite content, and citation frequency |
| Grounding Visibility | Methods for visualizing what AI bases its responses on |
| LLM Retrieval Optimization | Document optimization for the LLM retrieval layer |
| AI-native Documentation | Research on document formats that LLMs handle well |
| Agent-oriented Information Architecture | Information structures that AI agents can operate on |
Why This Initiative
Section titled “Why This Initiative”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.
Research Principles
Section titled “Research Principles”| Principle | Meaning |
|---|---|
| Reproducibility first | Every metric ships with a calculation formula and an OSS checker |
| Draft over Standard | Specifications are published as “Draft / Experimental / Proposal v0.1” so they remain revisable |
| Open Source first | Tools under OSS licenses, data under CC BY, specifications under MIT |
| Solo-honest | Solo operation is stated explicitly rather than dressed up as a consortium |
Founder
Section titled “Founder”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:
- Books: Full book list (kenimoto.dev/books)
- LLMO series (Kindle / Zenn Book, in Japanese, English, Portuguese, and Spanish)
- Harness engineering series (Kindle / Zenn Book)
- Web: kenimoto.dev / propel-lab.co.jp
- Amazon Author Page: Ken Imoto on Amazon
- Zenn: zenn.dev/kenimo49
- OSS: github.com/kenimo49
Phase Roadmap
Section titled “Phase Roadmap”The Initiative matures in phases. Each phase is the prerequisite for the next.
| Phase | Scope | Status |
|---|---|---|
| Phase 0 | Research framing, Mission publication, first Experiment Log | In progress |
| Phase 1 | Reproducibility — OSS CLI (llmo-checker), Score v0.1 Draft, dataset publication | Planned |
| Phase 2 | Community — contributors, external references, feedback channels | Planned |
| Phase 3 | Standardization — formal specifications, Compatible certification badge, Working Group formation | Planned |
Standardization comes last. Without mature OSS, benchmarks, and implementations to back them up, neither certification nor specifications can earn trust.
Contribute
Section titled “Contribute”| Method | Link |
|---|---|
| Issues / bug reports | github.com/kenimo49/llmo-guide/issues |
| Pull Requests | github.com/kenimo49/llmo-guide |
License
Section titled “License”This site and all draft specifications are published under the MIT License.