Discovery access
HTTP status, HTTPS, indexability, canonical routing, mobile viewport, robots.txt, and sitemap.xml.
Public-signal audit
See what answer engines can actually read.
Suede inspects public HTML, crawler rules, entity schema, sitemap coverage, metadata, and content structure. Every failed check points to a repair.
Check your inbox. We emailed a 6-digit code.
The code expires in 15 minutes. Nothing arriving? Check spam, or with a different address.
Inspecting public site signalsChecking the homepage and discovery files.
Report preview Live evidence
yourdomain.com
Public-site discovery and answer-readiness analysis
Readiness scoreCalculated from weighted public checks
01Crawler accessrobots.txt
02Entity evidenceJSON-LD
03Content clarityHTML
Every result links an observed signal to a repair.
How it works
Three steps. No account.
-
01
Enter a public URL
Give the audit any public homepage. Nothing is crawled behind a login, and no account is created.
-
02
Suede inspects the evidence
It fetches the homepage and root discovery files, then reads crawler policy, entity schema, and content structure.
-
03
Leave with repairs
Every failed check shows the value we observed and the specific repair that clears it.
What it checks
A report built from inspectable signals.
Suede separates public-site readiness from actual AI recommendations. The score covers what the audit can verify.
Access + Technical
Answer-engine crawl policy
Homepage access for OAI-SearchBot, Claude-SearchBot, PerplexityBot, and Googlebot, with the exact policy state shown.
Entity and content evidence
JSON-LD, primary entity type, sameAs links, headings, section length against the retrieval chunk limit, internal evidence links, page depth, and social metadata.
Why this differs
A checklist guesses. An audit observes.
Generic advice assumes a state. Suede reports the state it measured, then names the repair.
| Generic SEO checklist | Suede Audit |
|---|---|
| Tells you to add schema. | Shows the JSON-LD it parsed and the primary entity type it found. |
| Assumes answer engines can reach you. | Reports the robots.txt policy state for each named answer-engine crawler. |
| Scores every item equally. | Weights blockers above polish, so a blocked crawler outranks a missing preview image. |
| Ends at a list of problems. | Ends at one repair per failed signal, with the observed value attached. |
Example report
What you get back.
A shortened example of the report layout. Run the audit to see your own measured values.
Sample dataIllustration only. Not a live audit, and not a real domain.
example-saas.com
Public-site discovery and answer-readiness report
68/ 100
Grade C
Discovery access82
Crawl policy41
Entity evidence74
Crawler access
- GooglebotAllowed
- OAI-SearchBotAllowed
- Claude-SearchBotBlocked in robots.txt
- PerplexityBotBlocked in robots.txt
Prioritized repairs
-
Claude-SearchBot and PerplexityBot are disallowed.
Observed:
Disallow: /under both user-agents. Repair: remove those two disallow rules from robots.txt. -
No Organization JSON-LD on the homepage.
Observed: no
@type: Organizationnode. Repair: add Organization schema with name, url, and sameAs links. -
sitemap.xml is not referenced in robots.txt.
Observed: sitemap reachable, no
Sitemap:line. Repair: add the absolute sitemap URL to robots.txt.
Method before score
Truth stays attached to the artifact.
The audit fetches one public page plus root discovery files. It scores weighted checks, shows the measured value, and names the repair for every failed signal.
A high score is not proof of an AI citation. Buyer-prompt evidence requires separate, timestamped runs inside each answer engine.
Read the audit source Read the full scoring methodQuestions
What the free audit actually checks.
What does the free audit check?
The audit inspects your public homepage plus root discovery files: robots.txt, llms.txt, and sitemap.xml. It scores crawler access, entity and content evidence, and answer-engine crawl policy, then names a repair for every failed signal.
Does a high score mean AI engines will cite my site?
No. A high score is not proof of an AI citation. Buyer-prompt evidence requires separate, timestamped runs inside each answer engine, which this audit does not run.
How many free audits can I run?
One free audit per browser, on a public website. No account is created — you confirm a working email with a 6-digit code and the report unlocks on screen.
Why do you ask for my email?
The report unlocks with a one-time code sent to your address, which proves the email is real. We use it to send that code and to know who ran the audit. It is not stored in a database, added to a mailing list, or shared.
What happens after I get my results?
The report shows a weighted readiness score, a prioritized repair list, and a full findings ledger. Suede SEO can package the repairs as implementation-ready fixes or ship the edits with before-and-after evidence.
Who built this
The method is public because it should be checkable.
Suede Audit is built by Jason Colapietro, founder of Suede Labs AI. The scoring code is open source, so any result on this page can be traced to the check that produced it.
If a score looks wrong, read the source and tell us. That is a better contract than a number you cannot inspect.
- Open sourceEvery weighted check in this audit is published in the public repository.
- 21 merged pull requestsAcross 19 open-source projects, including Backstage and Jest.
- Suede Labs AIFounded 2024. Admitted to Stripe Startups in 2026.
Talk to a person
Questions about a score, or work you want quoted?
Send the domain you audited and the check you are asking about. Mail goes to the Suede Labs inbox.
Run it
See what answer engines read on your site.
One free audit per browser. Public websites only. No signup, no credit card — a verified email unlocks your report.
Run your free auditLive reportFresh automated audit
domain.example
Public-site discovery and answer-readiness report
0Weighted checks
0Audit lanes
0Repairs found
Readiness score
Three pillars. One public evidence trail.
The score weights blockers above polish. A blocked crawler or missing entity outranks a missing social preview.
Crawler access
Who can reach the public homepage.
These cards read robots.txt policy. They do not measure recommendations, rankings, or model answers.
Training and AI-use controls
Separate from search access.
GPTBot, ClaudeBot, and Google-Extended control model training or AI use. Blocking them does not block OAI-SearchBot, Claude-SearchBot, or Googlebot from search discovery, so these cards do not change the readiness score.
Answer Readiness Field
Every lane, plotted by measured readiness.
Rows move from weak signal to strong signal. High-impact blockers stay visible beside the score.
Weak signalStrong signal
Prioritized repairs
Fix the blockers before the polish.
Priority comes from the audit's weighted severity, not from a generic checklist order.
Full findings ledger
Every check, observed value, and priority.
| Status | Signal | Observed | Priority |
|---|
Methodology