VALIDATION PRICING · fixed scope · one site · no subscriptionreport: $49 · implementation: $199 · checks: 20
For small B2B SaaS, API and MCP product sites

Make your product legible to AI agents.

Your product may be clear to a person and still be awkward for software to discover or parse. I audit one public site against a transparent 20-check set, show the evidence, and either hand you the implementation plan or make the changes in your repository. This improves technical readability; it does not promise rankings, model mentions or citations.

Request the $49 report → View an example report Run the free score
report$49one public site · one business day
implementation$199one repository · up to five templates
current audit20published checks, weights and remedies
subscription$0the files stay in your repository
01 / fixed offers

Choose the answer, or choose the implementation.

Both offers cover one production website and use the same public 20-check audit. If the site does not fit the scope, I will say so before payment.

$49

Agent-readability report

For a team that will implement the work.

  • One public origin and one representative commercial page
  • Current 20-check result with request evidence
  • Prioritized fixes and implementation-ready snippets
  • Human review separating useful work from checklist theatre
  • Markdown and JSON delivery within one business day

Not included: repository changes, deployment, new marketing copy or third-party dashboard work.

Read the complete fictional example to see the evidence, priorities, owner inputs and implementation guidance before ordering.

Request the report
$199

Done-for-you implementation

For a team that wants the changes shipped.

  • One public site and one code repository
  • The report plus implementation on up to five core templates
  • Appropriate machine-readable files, metadata and discovery surfaces
  • One pull request or deployment, one revision and final re-audit
  • Delivery within seven calendar days after access

Not included: redesign, ongoing monitoring, bulk content creation, platform fees or inaccessible dashboard changes.

Request implementation
before orderingSend the public URL. For implementation, name the repository or platform. I confirm fit and delivery date.
available nowThe buttons open a pre-addressed email. I confirm the site fits, then send payment instructions; work starts after payment.
after deliveryRe-run the free score at index.percall.dev. Nothing is locked to me.
02 / implementation surface

The smallest coherent set of changes.

A product does not become more useful because it has every fashionable manifest. The report identifies the real gaps; the implementation adds only what describes an actual capability.

content_discovery

llms.txt, llms-full.txt, sitemap alignment, canonical URLs and useful machine-readable alternates.

typed_facts

schema.org JSON-LD with stable identities, explicit offers and claims that match the visible page.

crawler_policy

A deliberate robots.txt and Content Signal posture, including CDN rules the repository cannot override.

agent_surfaces

A2A card, Agent Skills, API catalog, MCP card or Markdown negotiation when the site really has those capabilities.

The target is technical accessibility, not a magic score. I will not publish a fake API catalog, an MCP card for a service that does not exist, or a signing directory without a signing key just to turn a check green. Owner-only configuration is reported separately.
03 / historical evidence

The original audit was tested on the whole fleet.

This is the July 28 result under the original 13-check v1 audit. The public default expanded to 20 checks in August; the numbers are retained as a historical case study, not presented as current scores.

v1 before / after · 12 sites · same 13 checksaverage 68.6 → 97.5
sitebeforeafter
index.percall.dev (then index.kc-it.pl)A 100A 100
kc-it.plE 48A 100
110kc3.github.ioC 79A 100
…/polish-sweepstakesA 93A 100
…/rentgen-ofertC 78A 100
…/bankier-street-betsC 73A 100
przetargimiejskie.plD 69A 100
…/puzleD 67A 100
…/pool-gliwice-automationD 60A 100
stareaparaty.comE 54A 90
protocolindex.euD 64A 90
overtimelog.comF 38A 90
Historical result only. The three v1 scores at 90 were held by CDN-managed crawler rules, not by a repository change.
What carried forward: origin-level signals should be fixed once, page metadata at the template, and CDN policy can contradict the repository. The current scorer, weights and remedies are public at index.percall.dev/checks/.
04 / fit

Useful plumbing, with a clear boundary.

Good fit

A small B2B SaaS, API or MCP product with a public production site and someone who can approve repository changes.

Not a fit

Guaranteed ChatGPT placement, reputation management, mass page generation, or a score purchased without changing what the site truthfully offers.

One URL is enough to start

Get the gaps, evidence and order of work.

The $49 report is for teams who will implement. The $199 offer is for teams who want the files and templates changed for them.

Request the $49 report → View the example report Request the $199 implementation