The Miatz Index scores how effectively a company educates and represents itself — to customers, developers, and now the AI agents reading on their behalf. Public materials only. Published methodology. Right of reply, always.
Four pillars are assessable from the outside and make up the public Index. The fifth needs the inside of the building, so it is deliberately kept out of the leaderboard.
Is there an academy at all, and does it work? Presence and reachability, content freshness, path coverage across the roles that actually use the product, and how rigorous the certification is — if certification exists.
How ready employees, support and customer-success teams are to explain the product they sell. Judging this needs access to internal material, so it is never scored from the outside and never appears in the public leaderboard.
API and developer-experience quality: reference completeness and spec sync, working sandboxes and keys, a real error catalog, changelogs and migration notes that a developer can act on.
Whether machines can use what you published: MCP endpoints, llms.txt and .md mirrors, answer-shaped structure, and how visible and citable the material is to answer engines.
Whether docs, academy, marketing and AI surfaces tell one story. Contradictory terminology, stale claims and three different versions of the same limit all cost points here.
A leaderboard is only worth reading if you can check it. These are the rules we hold ourselves to, and the ones you should hold us to.
Public docs, public academies, public APIs, public MCP endpoints. Nothing behind a login, nothing bought from a data broker, nothing scraped from a private surface — ever.
If a company tells crawlers to stay out, we stay out and record the pillar as not assessable rather than guessing. A blocked corpus is a finding, not an excuse to bypass.
The criteria, weights and rubric version publish alongside the numbers. Every score is dated and tied to a rubric version, so a company can see exactly what changed and why.
Internal enablement is assessed only inside a paid engagement with the company's consent and their own material. It never contributes to a public Index score.
Each criterion resolves to an observable artifact with a captured reference. Two assessors reading the same rubric should land within a point of each other; where they cannot, the criterion is broken and we fix the rubric.
The Index exists to raise the floor for how software companies teach. Findings are specific and fixable, framed as work orders — not as a ranking designed to embarrass anyone.
The rubric behind the Index is exactly the kind of judgment the miatz curriculum trains — reading a company's public surface and telling real signal from a facade.
Methodology v1 and the first scored cohort publish together. Until then this is the shape of the table, not a preview of anyone's number.
| # | Company | Index | Last scored |
|---|---|---|---|
No public scores yet The first cohort of scored companies publishes with Miatz Index methodology v1. Want to know when — or want your company assessed first? Tell us. | |||
Scores will always carry the rubric version and the date they were captured. An Index entry is a snapshot of public material on a given day, not a permanent verdict on a company.
A framework for evaluating how effectively a company educates and represents itself, across five pillars: customer education effectiveness, internal enablement, developer connectivity, AI and agent connectivity, and brand and knowledge representation consistency. The public leaderboard scores only the pillars that can be assessed from public materials.
The Docs Score grades one corpus — a documentation set against a 22-criterion rubric, with a per-page gap report. The Index sits above it and grades the company: docs are one input, alongside the academy, the developer surface, the agent surface and how consistent the whole story is.
No. Public docs, public academies, public APIs, public MCP endpoints, and nothing else. Internal enablement is the one pillar that needs internal material, which is exactly why it is excluded from public scoring and assessed only inside a paid engagement.
No. Buying an engagement gets you the private internal-enablement assessment and a fix queue; it does not move a public number. If a public score changes it is because the public materials changed, and the re-score is dated so anyone can see that.
Write to us. Every scored company gets the full criterion-level report with the captured references, a named contact to dispute any single criterion, and a re-score once a disputed item is fixed. If we got something wrong we correct it and say so on the entry.
Methodology v1 and the first scored cohort publish together — we are not shipping numbers before the rubric they came from is public. This page is live now so the method can be read, and argued with, before any company is on it.
The Miatz Docs Score is the corpus-level rubric underneath the Index — run it on your own documentation and get the gap report before anyone else reads it.