GlyphGuard

Unicode 17 · UTS #39

They look alike.
They aren’t.

GlyphGuard exposes deceptive characters before they reach a username, domain, moderation queue, or security log.

Inspect the API OpenAPI 3.1
6,565
confusable mappings
10k
code points / field
No fetch
deterministic analysis
Specimen 0042 Lowercase “p” lookalikes
p Latin U+0070
р Cyrillic U+0440
ρ Greek U+03C1
Visual verdict Similar shape. Different identity. confusable
Case fileGG–017
раypal.com

р U+0440 is Cyrillic, not Latin p.

а U+0430 is Cyrillic, not Latin a.

Risk100critical

The instrument panel

Ask five precise questions.

No generated prose and no probabilistic verdicts—just bounded Unicode analysis you can put in a pipeline.

POST/v1/analyze-textWhat is hidden inside this string? POST/v1/analyze-domainIs this domain impersonating one I trust? POST/v1/compareCould a person mistake these strings? POST/v1/sanitizeWhich unsafe controls can I remove? POST/v1/batchCan I inspect up to 50 items together?

One request, visible evidence

Keep the finding.
Skip the guesswork.

Responses include scripts, code points, normalized forms, security skeletons, and a bounded risk score. Findings are explicit enough to log, review, or act on.

Inspection transcriptapplication/json
POST /v1/compare
{
  "left":  "раypal",
  "right": "paypal"
}

HTTP 200
{
  "visually_confusable": true,
  "exact_match": false
}

Base mainnet · USDC · x402

$0.001

per single operation

Batch, up to 50 operations$0.004
Account requiredNo
API key requiredNo
SettlementUSDC on Base

Your client receives an ordinary HTTP 402 challenge, signs the payment, and retries the same request.

Build your first call