AI-search citation,
audited & governed.
ChatGPT, Perplexity, Google AI Overviews and Claude answer questions about your brand before a single user reaches your site. We read how the engines see you, fix the structured-data and entity signals that move citations, and write every change to a signed, append-only audit log your auditors can read.
Built for regulated and multilingual SMBs across Europe, the UK and Morocco. Six automated audit checks, every result reviewed by a human strategist. A signed governance pact between you and us. Every proposed change logged, reversible for 30 days, built with AI Act Article 50 transparency obligations in mind. The brands the engines cite are the ones that earned it - and can prove it.
The control plane for AI-search visibility.
Six capabilities, one product. Each is auditable, logged, and rolls back cleanly. No magic, no auto-publish, no surprises.
Six automated audit checks
Run on every sweep. AI agents are starting to transact, not just answer - so the bench scores how machines read you and whether they could act on you.
HTML / AEO legibility
Schema.org JSON-LD
PageSpeed
CrUX field data
AI citation rate
Agent readinessNEW
Signed governance pact
Per-site, version-pinned, sha256-gated. Every mutation lands in an append-only audit log with hashed actor IDs. Revocation is one endpoint away.
Knowledge graph
A growing knowledge graph of canonical entities. Schema.org JSON-LD per entity URL. Wikidata sameAs anchors. Cross-link engine proposes entity-justified edges between member sites.
Multilingual entity consistency
Hreflang reciprocity and entity coherence reviewed in EN, FR, AR and DA - by a human strategist who reads all four, not machine translation. Most auditors don't touch this. We treat it as foundational.
GDPR Article 17 erasure
A single endpoint tombstones a site's subject data while preserving the hashed event trail. Built with EU AI Act transparency obligations in mind - versioned governance, sha256-gated policies, append-only decision log.
Regression alerts
Per-source budgets with 10pt tolerance. Daily delta. budget_breach / regression / streak_recovery events. Idempotent dedup per period. Webhook delivery on the way.
Four steps. Each one writes to the record.
Nothing reaches your CMS without your approval. The architecture refuses to auto-publish.
Register your domain
We fetch your sitemap, parse robots.txt, detect the CMS type. Locale, currency, country noted. You're in the registry.
POST /sites { "tenant_id": "you", "domain": "yoursite.com" }
Sign the pact
Version-pinned governance pact, sha256-gated. Every mutation from here lands in an append-only audit log with a hashed actor ID.
POST /governance/pact/sign { "site_id": N }
Audit, then review
Six automated checks per sweep, reviewed by a human before anything ships. Scores feed a trend, budgets, and alerts. Fail-closed when credentials are missing. No silent zeros.
POST /audits/run · /audits/html · /audits/schema
Propose, you approve
Entity-justified cross-links, schema fixes, a back-of-folio list. Proposals land in a queue. Your infra pulls approved changes.
GET /sites/{id}/approved
Programmable, by design.
Every audit, every entity, every approval is reachable via a clean REST surface. HMAC-keyed, rate-limited, content-negotiated. api.citetome.com
# Probe an entity in the catalogue curl https://api.citetome.com/kg/entity/indexnow { "slug": "indexnow", "kind": "DefinedTerm", "name_en": "IndexNow", "wikidata_id": "Q108546024", "references": [ { "site": "sheandelle.com" } ] }
What the catalogue holds today.
Real counts from production, not slides. Updated on each audit sweep.
June 2026: we audited 38 Danish SMB websites.
Five industries, scored by the engine, reviewed by a human. Average AI-visibility 83 of 100. Average agent-readiness 48 of 100. Not one publishes the structured prices, availability or opening hours an AI agent needs to act. Anonymized aggregates only - we never name audited companies.
One brand is a node. The network is the product.
Citetome runs on a knowledge-graph engine that links member sites through shared entities. The more sites cite the same canonical entity, the more weight it carries in AI search. Public surfaces below.
Entity catalogue
Every canonical entity, addressable by stable @id. The surface AI engines crawl.
Posture & audit log
Live policies, append-only event log, GDPR Article 17 erasure endpoint. Public.
Customer console
Per-tenant operator dashboard. Audit history, approval queue, score trend.
Live 3D globe
Interactive portfolio view. See member sites, monitored links, pending approvals.
LLMs registry
The llms.txt file AI crawlers fetch first. What the network surfaces to them.
Fixed-fee. No metered surprises.
Pricing validated against the EU SMB market by an external advisor. Start with an audit, upgrade when you're ready.
- 11-signal HTML/AEO audit
- Strategist-reviewed audit report
- 5 working-day turnaround
- 15-min walkthrough call
- Everything in Audit
- Human multilingual review (EN/FR/AR/DA)
- Cross-link opportunity map
- Entity catalogue + 10-fix backlog
- Monthly audit sweep
- 1 governance review
- Monthly visibility report - engine-drafted, human-reviewed
- Registry membership
- Everything in Starter
- Entity onboarding
- Hands-on CMS implementation
- Monthly cross-link proposals
- Priority email support
- Everything in Standard
- Human-approved CMS edits
- Multilingual hub coordination
- Quarterly strategy call
Show me how my site reads to AI.
Send your URL on the audit page. We run the automated AEO + GEO checks, review the findings by hand, and send the report within five working days. Optional 15-minute walkthrough on request. No commitment, no upsell.
Request an audit →Common questions.
What does Citetome actually do?
We audit your website for AI-search readability, fix the structured-data and entity signals that ChatGPT, Perplexity, Google AI Overviews and Claude rely on for citation, and run a governance trail so every change is signed, audit-logged and reversible.
How is this different from regular SEO?
Regular SEO optimises for keyword ranking. Citetome optimises for entity legibility - whether AI search systems can identify, cite and trust your brand. Both matter; the AI-search citation layer is what's new and unowned for SMBs in 2026.
Will you write to my CMS automatically?
No. Architectural rule. Proposals land in an approval queue. A human approves. Your own infrastructure pulls approved changes. We never auto-publish.
Why does the governance trail matter?
EU AI Act enforcement begins August 2026. Regulated SMBs - legal, finance, healthtech - need an auditable record of every change made to their AI-search visibility. We sign a pact with each member site, write every proposal and approval to an append-only log, and support GDPR Article 17 erasure on demand.
AEO or GEO? Which one is this?
Both. AEO (answer-engine optimization) covers featured snippets, voice search, and Google AI Overviews - the "extract a direct answer" layer. GEO (generative-engine optimization) covers ChatGPT, Perplexity, Claude, Gemini, Copilot - the "synthesize and cite a brand" layer. Citetome audits, fixes, and governs both, because in 2026 the same SMB needs the same brand to be readable, citeable, and trustworthy across every machine reader.
Who's behind it?
Solo-founded in Copenhagen, built on Fly.io with a Python + FastAPI + SQLModel stack. 197 automated tests passing. Pricing validated by an external advisor with European SaaS experience. Markets served: Denmark, the UK, Morocco, and any multilingual SMB with the same problem.
Five days. One audit. Zero commitment.
If the catalogue is for you, we sign the pact and continue. If not, you keep the report.