Canonry Custom · Scoped engagements
Your agency's visibility infrastructure.
A dedicated Canonry application and data layer joins AI answer evidence, Google Search Console, GA4, Bing, server traffic, and connected ChatGPT Ads accounts where beta access is available.
Scope hundreds of markets under one site, then give your team a purpose-built application UI and its agents the right project-scoped tools.
Evidence, with the collection method attached
A scrape is one observation.
A consumer interface can vary by account, geography, history, subscription, routing, model, and time. One session should not become an unexplained market-wide score. Every AI Visibility Tool Is Lying to You sets out that measurement problem in full.
Canonry uses APIs for controlled measurement and labels optional browser evidence when the interface itself is under test.
More than AI visibility
Evidence across visibility, traffic, and outcomes.
Canonry turns connected evidence into one project-scoped operating record your team can query, report, and act on.
AI visibility in business context
Every signal keeps its source.
Connect only what the engagement needs. Canonry preserves where each observation came from and how it can be used.
- 01
AI answers
Mentions, citations, competitors, sources, prompts, and run history
- 02
Search
Google Search Console, Bing, and optional Google Business Profile evidence
- 03
Engagement
GA4 traffic, AI referrals, acquisition, configured lead events, and landing-page behavior
- 04
Server traffic
Requests attributed to AI crawlers, user fetches, and referrals in supported logs
- 05
Paid
Available ChatGPT Ads account state, performance, configured conversion events, and operation receipts
- 06
Technical
Audits, schema, indexing, backlinks, content opportunities, and approved actions
Custom measurement model
One domain can contain hundreds of markets.
A domain-level score can hide the parts of a business that must be measured separately. Canonry can scope locations, properties, brands, services, audiences, competitors, and local query sets as distinct targets. The write-up on one site with many markets walks through targets, URL coverage, and shared queries.
A separate instance protects each unrelated client trust boundary. Within an instance, project-scoped access limits operators to the agreed tools and evidence. Shared instance settings remain shared.
Server-side traffic
See server-observed requests attributed to AI systems.
On supported hosts, Canonry reads the request logs already produced by the platform serving the site. Cloud Run and Vercel require no Canonry browser tag or changes to the served site's application code. Canonry records attribution and verification status where the source data permits.
WordPress uses a prebuilt Canonry logging plugin. During scoping, we document source coverage, cache limitations, authentication, and retention before calling the data complete.
Which client markets lost visibility and traffic after the last content release?
visibility.compare · gsc.performance · ga4.traffic · traffic.eventsTwo markets declined across both citations and organic sessions. The raw answers, affected pages, Search Console queries, and server events are attached.
Built for the agent you already use
A custom application for people. Tools for agents.
Every Custom deployment ships a purpose-built Canonry application UI for people who run the work. Canonry also exposes typed CLI, REST, MCP, webhook, and native agent-plugin surfaces, the same ones documented on the open-source Canonry platform. A custom engagement selects the integrations and tool coverage it supports. A ChatGPT connector can be separately scoped and validated.
Agent-first means the operating record is machine-actionable. It does not remove human control. Access, approvals, credentials, and spend-bearing actions stay bounded by the deployment.
Custom deployment flow
Start with the operating model.
Canonry Custom begins with what must stay separate, which evidence already exists, and how the application and integrations need to support the work.
- 01
Model the business
Define the clients, markets, properties, brands, audiences, competitors, queries, and permissions that must stay separate.
- 02
Connect the evidence
Connect the relevant AI providers, search platforms, analytics, hosting logs, technical sources, and Ads accounts where beta access is available.
- 03
Ship the application where work happens
Canonry ships a purpose-built application UI for the agreed workflow, then exposes the same project-scoped evidence through APIs, agents, client reporting, webhooks, and agreed connectors.
Direct answers
Before we scope the deployment.
What is Canonry Custom?
Canonry Custom is a scoped engagement for agencies and complex organizations that need a dedicated Canonry deployment. Canonry models the business structure, connects the relevant evidence sources, and ships a purpose-built application UI alongside the agreed APIs, agent tools, reports, webhooks, and connectors.
Does Canonry Custom include an application UI?
Yes. Every Custom deployment ships a purpose-built Canonry application UI for the agreed operating workflow. The UI does not replace agent or API surfaces. They use the same project-scoped evidence and permissions.
How is this different from an AI visibility dashboard?
A mentions-and-citations dashboard reports what appeared in AI answers. Canonry can join that evidence with Google Search Console, Google Analytics 4, Bing Webmaster Tools, supported server logs, technical evidence, and available ChatGPT Ads data. The result is one operating record for investigation, reporting, and approved action.
Does server-side traffic require changes to the website?
Not on every supported path. Canonry can read existing Cloud Run and Vercel platform logs without a new browser tag or changes to the served site's application code. WordPress uses the prebuilt Canonry Traffic Logger plugin. Source availability and cache limitations are confirmed during scoping.
Does Canonry only use APIs?
Canonry uses provider APIs for controlled, repeatable observations. Optional browser collection is available when the consumer interface itself is the measurement surface. Canonry labels the provider, model, context, location, collection surface, timestamp, answer, and citations so API and browser observations are never presented as interchangeable.
Can our team use Canonry from ChatGPT, Codex, or Claude?
Canonry already exposes CLI, REST, MCP, webhooks, and native Codex and Claude Code plugins. Available tool coverage varies by integration and is confirmed during scoping. A ChatGPT connector can be evaluated as separate integration work, subject to technical validation.
Does agent-first mean the agent can change anything?
No. Agent-first means Canonry evidence and workflows are machine-actionable. Separate instances protect unrelated client trust boundaries. Within an instance, project access, provider credentials, write authority, approvals, and spend-bearing actions are scoped by the deployment.
Canonry Custom
Build the visibility service your agency needs.
Tell us what must be measured separately, which systems hold the evidence, and where your team and clients need the application and its connected tools.