AI evaluation and verification engine

The engine behind AIOInsights.

AIOTruth is the evaluation and verification engine behind AIOInsights. It measures the public signals a business publishes, shows the evidence behind each finding, and refuses to score what it cannot observe.

Deterministic scoring. Evidence shown. Nothing invented.

Run the check on AIOInsights How the engine works

What the engine evaluates

Can AI systems access, understand, verify, and represent this business from the evidence it publishes?

Access

Whether crawlers and agents can reach the pages, data, and declarations at all.

Understand

Whether identity, entity, and meaning are consistent and machine legible.

Verify

Whether claims are attributable, sourced, and corroborated rather than asserted.

Represent

Whether an AI system can describe the business accurately from what it finds.

Evaluation types

Each type runs behind an AIOInsights check and is labeled Powered by AIOTruth. The check happens on AIOInsights; AIOTruth is the methodology and the evidence.

Evaluation typeLive

Trust Visibility

Evaluates the observable signals a business publishes across clarity, entity consistency, authority, AI discoverability, trust, and local presence.

Content Credibility

Evaluates whether a page is readable, attributable, sourced, specific, original, and honestly maintained. It does not claim to prove whether a human or an AI wrote the page.

WebMCP Readiness

Inspects the current evidence that a website exposes structured tools for browser agents. Absence is not a current failure, and presence is not a security certification.

All evaluation types

Deterministic and evidence based

Every point comes from an observable signal. The same input scores the same result twice, and each finding shows the evidence behind it.

Methodology

Inputs, retrieval, signals, weighting, failure states, and what each type does not measure.

Changelog

Engine v2.4.0. Every version that produced a result, and what changed.

The engine

What AIOTruth is, what it owns, and how it powers AIOInsights.

Run the check, then act on it

AIOTruth introduces the finding. AIOInsights runs the check. Digilu implements and manages the work.

Run the check on AIOInsights Ask Digilu to implement the findings

Compare with AIOStory. Define a term with AIOFacts. Watch the environment with AIOWeather.

AIOTruth Research

The AIO worth knowing.

The evidence labeled record of the tools, platforms, methods, experiments, and claims shaping the field. This is the research layer, scored by the Research Value Score, and separate from the evaluation the engine runs for a specific business.

Best Today

Scored 8.0 or higher, published or found in the last 72 hours, and never Unverified.

Models & Platforms 8.1

2 months (meme-d) research, Do People Notice the Difference Between AI Models?

If ordinary users cannot tell which model is answering, then the brand name on an AI interface is carrying much of the trust, not the model underneath.

Supportedhuggingface.co

Models & Platforms 8.1

Aplomb 1: open-weights 5.3B decision model, 1M context, text/image/video/audio in one request, #1 among 4B models on the Decision Index

Decision models like this sit in the routing layer of AI agents, where a system chooses which tool to call, which document answers the question, and whether the input contains an answer at a

Confirmeddocs.empiriolabs.ai

Tools & Apps 8.4

Agentic AutoRAG: RAG Pipeline Optimization through Reasoning-Driven Agents

Most AI answers about a brand come out of a retrieval pipeline, and this work treats the split between retrieval failure and generation failure as the central diagnostic.

Supportedarxiv.org

Best This Week

The strongest items of the current weekly period, ranked by Research Value Score.

1

What can Claude Code really do for GEO? 30 skills tested

Models & PlatformsSupported8.7/10

A score produced inside a coding skill is easy to mistake for a measurement of how AI systems treat a page.
2

Agentic AutoRAG: RAG Pipeline Optimization through Reasoning-Driven Agents

Tools & AppsSupported8.4/10

Most AI answers about a brand come out of a retrieval pipeline, and this work treats the split between retrieval failure and generation failure as the central diagnostic.

Read the full weekly roundup

Tools & Apps

Tools & Apps 8.4

Agentic AutoRAG: RAG Pipeline Optimization through Reasoning-Driven Agents

Most AI answers about a brand come out of a retrieval pipeline, and this work treats the split between retrieval failure and generation failure as the central diagnostic.

Supportedarxiv.org

Tools & Apps 7.5

Explain how an importance gate prevents noise accumulation in persistent agent memories.

What an AI agent remembers about a brand depends on what its memory layer agrees to write, not only on what the agent reads.

Confirmedmnemoverse.com

Tools & Apps 7.9

Free ChatGPT Rank Tracker | Track AI Citations

For a brand trying to be recommended by AI, the practical question is whether ChatGPT cites it for the prompts buyers actually use and where it sits among the sources it cites.

Supportedapiserpent.com

Tools & Apps 7.5

I post-trained Qwen3.6-35B-A3B into my daily-driver local coding/agent model QwiVer3.6-35B-A3B GGUF

The set of models reading, summarizing and recommending your content is no longer a short list of hosted assistants.

Supportedhuggingface.co

All Tools & Apps

Models & Platforms

Models & Platforms 7.5

Top 10 Lists and AI Search: What Still Works in 2026

This touches how brands are found through AI and citation, recommendation, and representation.

Disputedahrefs.com Models & Platforms 7.6

How to Get Your Brand Discovered by ChatGPT & AI Search | CitableHub

A brand that wants to be cited by AI assistants needs its facts in a form machines can parse, and this directory offers a ready-made structured profile, feed, and agent endpoint without the

Supportedcitablehub.com

Models & Platforms 7.6

Free AI Visibility Checker: See Where Your Brand Ranks in AI Answers

Most brands have no baseline for how AI systems describe or rank them, so they cannot tell whether a change helped.

Supportedtrackmyvisibility.com

Models & Platforms 8.1

2 months (meme-d) research, Do People Notice the Difference Between AI Models?

If ordinary users cannot tell which model is answering, then the brand name on an AI interface is carrying much of the trust, not the model underneath.

Supportedhuggingface.co

All Models & Platforms

Content & Methods

Content & Methods 7.5

What AEO ranking means for your visibility in AI answers

The guide frames AI visibility as an evidence problem: a business gets named in an AI answer when the engine can read and cite clear, verifiable facts about that entity.

Supportedtherankingfactory.com