SYNTHIC

Platform

Everything between your brand and the answer

Six capabilities, one loop: measure what the engines say, protect what's true, optimize what they read — then measure again.

Start free audit View pricing
Northwind Coffee · visibilitySample

Visibility score

58

+12 vs. first audit

Mention rate

53%

240 answers scored

Citation rate

31%

114 links returned

Visibility · weekly

12 wk

Per engine · mention rate

  • ChatGPT71%
  • Claude58%
  • Gemini46%
  • Perplexity63%
  • Grok29%

Share of voice · who gets named

  • Blue Bottle186
  • Northwind CoffeeYOU139
  • Stumptown97
  • Counter Culture58

Connected via official engine APIs

ChatGPTClaudeGeminiPerplexityGrok

01MONITOR

AI Search Audits

Your buyers' questions, sent to every engine through its official API. Each answer stored verbatim, then scored: named, how prominently, in what tone.

Deep dive
Northwind Coffee · visibilitySample

Visibility score

58

+12 vs. first audit

Mention rate

53%

240 answers scored

Citation rate

31%

114 links returned

Visibility · weekly

12 wk

Per engine · mention rate

  • ChatGPT71%
  • Claude58%
  • Gemini46%
  • Perplexity63%
  • Grok29%

Share of voice · who gets named

  • Blue Bottle186
  • Northwind CoffeeYOU139
  • Stumptown97
  • Counter Culture58
FIG.01 — the visibility board · sample data, illustrativeMONITOR
  • ChatGPT, Claude, Gemini, Perplexity, Grok
  • Every score traces to a stored answer
  • Re-run and diff over time

02PROTECT

Hallucination Defense

Synthic crawls your site for ground truth, asks each engine about you, and checks every claim against what you actually say.

Deep dive
Hallucination scan · claims about youSample

Claims found

24

Inaccurate

6

corrections ready

Unverifiable

5

no page says either way

  • “Synthic starts at $99/month.”

    → Actual: free first audit, Founder at $49/mo

    Inaccurate

  • “Supports ChatGPT, Claude, Gemini, Perplexity, Grok.”

    ✓ Matches the site

    Accurate

  • “Founded in San Francisco.”

    Site is silent on location

    Unverifiable

  • “No longer offers schema generation.”

    → Schema & llms.txt are live features

    Inaccurate
FIG.02 — claim → verdict, with the correction · illustrativePROTECT
  • Verdict per claim: accurate, inaccurate, unverifiable
  • Severity so the damaging ones go first
  • Correction written from your own pages

03OPTIMIZE

Schema & llms.txt

Point it at a page and get the structured layer engines parse best — JSON-LD, llms.txt, answer-first FAQ — built only from what the page says.

Deep dive
AEO audit · /pricingSample

Page score

64

of 90

Passing

7

of 11 checks

Fixes ready

4

copy-paste code

    Checks

  • AI crawlers can read the site14/14
  • llms.txt present0/10
  • sitemap.xml present6/6
  • JSON-LD structured data16/16
  • FAQ schema for answer retrieval0/8
  • Descriptive <title>6/6

Generated from your own pages

{
  "@type": "Organization",
  "name": "‹from the page›",
  "offers": […],
  "sameAs": […]
}

— llms.txt · site root
— FAQ · answer-first pairs
— entities · crawled content only

FIG.03 — page → score → structured layer · illustrativeOPTIMIZE
  • JSON-LD ready to paste before </head>
  • llms.txt for the site root
  • Nothing invented: crawled content only

04ACCESS

AI Crawler Control

Your live robots.txt, evaluated for the sixteen crawlers that feed the answer engines. A blanket rule that blocks them by accident shows up in seconds.

Deep dive
Crawler access · robots.txtSample

Agents tracked

16

Allowed

14

Blocked

2

answering without you

GPTBot
OAI-SearchBot
ChatGPT-User
ClaudeBot
Claude-SearchBot
Claude-User
PerplexityBot
Perplexity-User
Google-Extended
Googlebot
Bingbot
Amazonbot
Applebot-Extended
Meta-ExternalAgent
Bytespider
CCBot

robots.txt · the two lines that matter

User-agent: CCBot
Disallow: /          ← blocks a live answer engine

User-agent: GPTBot
Allow: /
FIG.04 — agents · red = blocked · illustrativeACCESS
  • 16 crawlers, the exact rule that matched
  • llms.txt and sitemap.xml checks
  • Server logs: who actually came

05COMPETE

Share of Voice

Every answer that skips you names someone else. Audits tally who, per engine and per prompt, with each mention linked to the answer it came from.

Deep dive
who the engines named · category promptsrank
brand A#1
brand B#2
you#3
brand D#4
brand E#5
FIG.05 — who the engines named · illustrativeCOMPETE
  • Declared rivals plus the ones engines surface
  • Per-engine leaderboard
  • Every mention links to its verbatim answer

06TARGET

Persona Simulation

The founder and the CFO don't get the same answer. Run the audit through five buyer lenses and see where you win, and with whom.

Deep dive
FIG.06 — one keyword, five lenses · illustrativeTARGET
  • Five personas built in
  • Persona-shaped prompts per keyword
  • Compare across personas over time

Under the hood

The parts you can inspect.

How a score is made, drawn from the code that makes it. If the weights change there, they change here.

FIG.07 — the evidence chain · every number walks back to a stored answerLIB/SCORING.TS
mention×55

named at all

prominence×30

first, listed, buried

citation×15

your page as source

FIG.08 — visibility score = mention×55 + prominence×30 + citation×15 · value illustrativeLIB/SCORING.TS
11 checks · 90 weight points → score out of 100lib/aeo.ts

Access · 30

  • AI crawlers can read the site
  • llms.txt present
  • sitemap.xml present

Structured · 24

  • JSON-LD structured data
  • FAQ schema for answer retrieval

Meta · 15

  • Descriptive <title>
  • Meta description
  • Canonical + Open Graph

Content · 21

  • Clear heading hierarchy
  • Answer-first / Q&A structure
  • Substantive content
FIG.09 — AEO site audit · the checks and their weightsLIB/AEO.TS

See it on your own brand.

The first audit takes minutes and runs on live engine APIs from the start.

Start free auditSee pricing
  • Free first audit
  • No credit card
  • Live engine data
  • Your data stays yours
  • Cancel anytime