Free tools Windows power users keep installed
One-click scans. No signup required.
iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Promptway’s is-agentic scan returned 74 out of 100 in a report published on August 21, 2026. The score and findings below are the author’s account of that scan, not a fresh or independently verified audit. The report’s useful lesson is as much about what not to change as what to fix: improve agent access where it serves the site, but don’t invent capabilities just to earn points.
What Promptway’s 74-point score means
In his August 21, 2026 field report, Promptway author Agnel Nieves said he ran npx is-agentic promptway.com and received a score of 74, labeled “Ready with a few material gaps.” He reported 59 of 80 Essential points, 12.6 of 20 Recommended points, and a 2.4-point bonus. These are the figures in that report; they do not establish Promptway’s score today. Read the original Promptway report.
![Claude AI for Beginners Bible: [5 in 1] The Ultimate Guide to Automate Your Work, Save Hours Every Week, and Use AI for Real-World Results](https://m.media-amazon.com/images/I/415+fSJacsL._SL160_.jpg)
The official is-agentic methodology describes a score based on how readily AI agents can discover, access, understand, and use a public website. Essential checks share an 80-point pool, Recommended checks share 20 points, and Emerging signals can add a bonus capped at five points. Checks deemed inapplicable are excluded, partial results earn proportional credit, and checks for surfaces such as APIs, OAuth, MCP, GraphQL, or MCP Apps activate only when that kind of surface is identified.
That makes the individual check results more actionable than the headline number. A publication and a public API are not the same kind of site, so a recommendation to expose an API should not automatically become a feature request for a magazine. The methodology also cautions that automated scans are point-in-time observations that can produce false positives, false negatives, or partial results. It explicitly says a high score is not a security, accessibility, quality, compliance, or compatibility certification.
#1 Best Overall
What the scan flagged
Nieves reported a mix of content, response-format, and site-identity gaps. Some findings concerned whether a useful destination existed; others concerned whether agents could recognize or interpret information that was already present.
- 404 recovery: The missing-page response already returned the correct 404 status, but its body offered no useful route forward.
- Homepage structure: The content available without JavaScript had a flat outline. Nested headings appeared inside card links and were not counted by the grader.
- Markdown negotiation: Requests with
Accept: text/markdownreceived HTML, and theVaryresponse header did not includeAccept. - Developer discovery: The grader could not find a page titled “Promptway developer resources.”
- Agent guidance:
llms.txtlisted pages but did not explain when Promptway was, or was not, a suitable source. - Trust and contact details: An About page existed, but Contact and Privacy pages did not.
- Organization data: The Organization JSON-LD lacked
contactPointandaddress. - Text-to-markup ratio: The report said readable text made up 4.83 percent of the homepage HTML, below the stated 5 percent threshold.
An MCP check also appeared because Promptway discussed MCP, even though the publication did not operate an MCP server. That is a useful distinction: a site can discuss a technology without providing it as a service.
What the author changed
Made missing pages recoverable
The report says the 404 page gained links to llms.txt, developer resources, the sitemap, archive, About, and Contact. The same missing URL was also made to return a short Markdown body when a request preferred text/markdown.
Added Markdown content negotiation
Nieves says Markdown negotiation was added for the same site URLs, with Vary: Accept so caches can distinguish responses by the request’s accepted format. Next.js RSC requests were excluded, while ordinary browser requests continued to receive HTML. The change was intended to serve an agent-friendly format without replacing the normal browser experience.
Created a clear developer and agent entry point
Developer information was documented at /for-agents under the title “Promptway developer resources,” alongside /openapi.json and /.well-known/api-catalog. The report also says llms.txt gained a “When to use Promptway” section.
That guidance named suitable jobs: AEO guidance for a live Next.js site, a prompt structure with a copy-and-paste example, a founder interview with a number, a tool write-up that says what to skip, and a sourced news filter. It also drew boundaries: Promptway was not the right source for a model API, MCP server, OAuth, webhooks, or unlabeled AI copy.
Filled in contact, privacy, and organization information
The author reports adding Contact and Privacy pages, an email address, and a San Juan, Puerto Rico location in JSON-LD. Organization schema was expanded with url, sameAs, logo, email, address, and contactPoint.
Quick wins for a faster PC:
Repair Windows errors before they cause bigger problemsFix Now →Fix the driver behind crashes, sound loss and screen glitchesFind Drivers →Clear out junk files and repair common Windows errorsFree Scan →Improved the server-rendered homepage outline
A “How to read this site” section added non-linked H2 and H3 headings, along with more prose available in server-rendered HTML. The stated approach was to add explanatory content rather than remove the magazine layout to improve the text-to-markup ratio.
What the author chose not to build
Promptway did not stand up an MCP server or publish a pretend manifest just to satisfy a check. It was a publication, not an MCP service. Likewise, the author did not strip out the magazine design to chase a better content-to-markup ratio. Instead, he added explanatory text that could help a reader or agent understand the site.
Those choices are the point of an applicability-aware audit: treat a score as a map of observed friction, not a mandate to add every possible interface. A recommendation is worth implementing when it supports a real capability or helps a genuine audience use the site. If it implies a service the site does not provide, creating a superficial version for a score can mislead users.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.How to reproduce and verify the checks
The report suggests running the CLI against the site and adding --json for structured output:
npx is-agentic promptway.com
npx is-agentic promptway.com --jsonFor any reported issue, verify what the site actually returns before changing it. For example, check the status on a nonexistent path and compare the response to a request that prefers Markdown:
Best Value
curl -i https://promptway.com/this-page-does-not-exist
curl -i -H 'Accept: text/markdown' https://promptway.com/this-page-does-not-existFor a deployed implementation, inspect the status, body, and relevant headers rather than assuming a code change worked. A Markdown response should reflect the request’s accepted format, and a response that varies by Accept should identify that variation in its cache behavior. Then deploy and rerun the scan to see whether the intended checks changed.
Timing matters when interpreting the result: the official methodology says a completed scan may be served from a six-hour freshness cache. Check the report timestamp and cache status where available; a result returned during that window may not represent a new scan. Nieves’s published 74 is a historical report, and the account does not establish a post-change score or independently verify Promptway’s current endpoints.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Do these 3 things before closing this tab:
1Clear out junk files and repair common Windows errors2Scan for outdated or missing drivers - takes under a minute3Repair Windows errors before they cause bigger problems
