Scanner Concordance, edition 2026-11-04
The first edition: what the scanners returned and where they disagree.
In short
Edition 2026-11-04, the first: sixty domains, four scanners that returned scores and two that gave no per-site lookup. Ora and Cloudflare rank the sites alike (Spearman 0.79) but sit far apart in level: medians 64.5 and 20. The widest spread is fastly.com at 73 points.
What was scored
| Scanner | Scored | Min | Median | Max |
|---|---|---|---|---|
| Ora | 60 of 60 | 27 | 64.5 | 100 |
| Cloudflare | 52 of 60 | 0 | 20 | 100 |
| is-agentic.com | 47 of 60 | 11 | 61 | 100 |
| agent-ready.dev | 0 of 60 | none | none | none |
| SiteSpeakAI | 0 of 60 | none | none | none |
| Mintlify | 21 of 60 | 39 | 93 | 100 |
How the scanners agree
| Pair | Shared domains | Spearman |
|---|---|---|
| Ora and Cloudflare | 52 | 0.79 |
| Ora and is-agentic.com | 47 | 0.80 |
| Ora and Mintlify | 21 | 0.47 |
| Cloudflare and is-agentic.com | 40 | 0.72 |
| Cloudflare and Mintlify | 18 | 0.39 |
| is-agentic.com and Mintlify | 21 | 0.42 |
A scanner that runs 20 points lower than another but orders sites the same way still correlates near 1. Ora and is-agentic.com correlate most strongly because one engine runs both; their mean gap is -2.4 points on 47 domains, and only 12 of those rows read the same stored scan, because is-agentic.com caches its own copy of an earlier Ora scan.
The ten biggest disagreements
| Domain | Ora | Cloudflare | Mintlify | is-agentic.com | Spread |
|---|---|---|---|---|---|
| fastly.com | 70 | 20 | 93 | 72 | 73 |
| jfrog.com | 70 | 27 | 98 | 56 | 71 |
| turso.tech | 84 | 33 | 99 | 92 | 66 |
| mux.com | 69 | 20 | 85 | 62 | 65 |
| llamaindex.ai | 71 | 20 | 81 | 53 | 61 |
| tavily.com | 92 | 40 | 99 | 89 | 59 |
| prisma.io | 86 | 40 | 96 | 97 | 56 |
| cockroachlabs.com | 69 | 20 | 48 | 56 | 49 |
| firecrawl.dev | 85 | 53 | 99 | 89 | 46 |
| airbnb.com | 37 | 20 | 59 | 27 | 39 |
Only 18 rows had three independent scores, so this list is mostly Mintlify scoring a documentation host against Cloudflare scoring the main domain. With agent-ready.dev and SiteSpeakAI giving no lookup, the disagreement table is thin; a later edition adds a keyed source if one is approved.
Ora against Cloudflare
| Domain | Ora | Cloudflare | Gap | Cloudflare level |
|---|---|---|---|---|
| dominos.com | 59 | 0 | 59 | L0 Not Ready |
| anakin.io | 88 | 31 | 57 | L1 Basic Web Presence |
| useboom.ai | 82 | 27 | 55 | L1 Basic Web Presence |
| temu.com | 66 | 13 | 53 | L0 Not Ready |
| tavily.com | 92 | 40 | 52 | L1 Basic Web Presence |
| llamaindex.ai | 71 | 20 | 51 | L1 Basic Web Presence |
| turso.tech | 84 | 33 | 51 | L1 Basic Web Presence |
| fastly.com | 70 | 20 | 50 | L1 Basic Web Presence |
| sorce.jobs | 70 | 20 | 50 | L1 Basic Web Presence |
| cockroachlabs.com | 69 | 20 | 49 | L1 Basic Web Presence |
Cloudflare's number here is the share of its counted checks a site passes, from a scan run at fetch time. Its level is gated: level 1 is basic web presence, level 2 adds AI bot rules and Content Signals, level 3 adds markdown negotiation, and level 4 adds one of MCP Server Card, A2A Agent Card, Agent Skills or API Catalog. A site can serve pages an agent reads well and still pass few of those checks, which is one reason the two numbers sit far apart while ranking the sites alike.
Reading the disagreements
| Scanner | Checks | Weighting and bands |
|---|---|---|
| Ora | Four layers an agent moves through: Discovery, Access, Usability, Payments; 132 checks in the API catalog, 126 on the methodology page | Discovery 20 points, Access 30, Usability 40, Payments 10; not-applicable checks leave the denominator. Grades A+ 95 to 100 down to F 0 to 27. |
| Cloudflare isitagentready.com | 23 checks in five categories: Discoverability, Content Accessibility, Bot Access Control, API, Auth and MCP Discovery, Commerce | Equal weights. Percentage passed of counted checks, and a level 0 to 5 gated by which groups pass. Commerce checks are shown and not counted. |
| is-agentic.com | Ora's checks, regrouped | Essential checks share 80 points, Recommended 20, a bonus capped at 5. Scored by Ora. |
| agent-ready.dev | 72 checks against the Vercel readability spec, llmstxt.org and agent protocol specs | Unweighted pass rate; a warning counts as a fail. Bands 90 and up excellent, 70 to 89 good, 50 to 69 fair. |
| SiteSpeakAI | Names WebMCP, llms.txt, structured data, content quality, accessibility and meta information | Publishes no check count, weights or bands. |
| Mintlify Agent Score | AFDocs checks on a documentation site; its pages report 35 checks | Weighted Critical 10, High 7, Medium 4, Low 2. Docs only. |
fastly.com: 73 points apart
Ora 70 (B), Cloudflare 20 (L1 Basic Web Presence), Mintlify 93 (A), is-agentic.com 72. Ora's layers: discovery 14 of 15, accessibility 42 of 56, usability 32 of 58. Cloudflare passed 3 of 15 counted checks, level 1. Mintlify scored www.fastly.com/documentation, not the main domain, so its score describes the documentation site alone.
jfrog.com: 71 points apart
Ora 70 (B), Cloudflare 27 (L1 Basic Web Presence), Mintlify 98 (A), is-agentic.com 56. Ora's layers: discovery 12 of 13, accessibility 42 of 57, usability 32 of 56. Cloudflare passed 4 of 15 counted checks, level 1. Mintlify scored docs.jfrog.com, not the main domain, so its score describes the documentation site alone.
turso.tech: 66 points apart
Ora 84 (B), Cloudflare 33 (L1 Basic Web Presence), Mintlify 99 (A), is-agentic.com 92. Ora's layers: discovery 12 of 14, accessibility 57 of 62, usability 79 of 101. Cloudflare passed 5 of 15 counted checks, level 1. Mintlify scored docs.turso.tech, not the main domain, so its score describes the documentation site alone.
Why a cell is empty
| Scanner | Reason | Domains |
|---|---|---|
| Cloudflare | scanner could not fetch the site (HTTP 403) | 7 |
| Cloudflare | scanner could not fetch the site (HTTP 404) | 1 |
| is-agentic.com | no completed report stored | 13 |
| agent-ready.dev | badge says unscored | 60 |
| SiteSpeakAI | web form only, not automated | 60 |
| Mintlify | no Mintlify score page | 38 |
| Mintlify | slug scores a different host | 1 |
The sample and its bias
Fifty of the sixty rows come from Ora's public leaderboard, which lists sites that someone scanned with Ora. That favours Ora and is-agentic.com coverage. The ten Tranco business sites are the control: none is on the leaderboard, and 58 of Ora's 60 readings came from its cache, some from scans run in September and October. Sampling, timing and access are on the method page; files are on the data page.
Sources
- Ora methodology page. https://ora.ai/methodology, read 2026-11-04.
- isitagentready.com. https://isitagentready.com/llms-full.txt, read 2026-11-04.
- agent-ready.dev methodology. https://agent-ready.dev/methodology, read 2026-11-04.
- Mintlify Agent Score. https://www.mintlify.com/score, read 2026-11-04.
- afdocs.dev. https://afdocs.dev/agent-score-calculation.md, read 2026-11-04.
- is-agentic methodology. https://is-agentic.com/methodology, read 2026-11-04.
Last reviewed 2026-11-05