Privacy

Simple public pages, clear boundaries, and cookieless analytics you can turn off.

Tonywood.org is designed as a public reading and agent-discovery site. It does not ask visitors to log in and does not expose private data through the public MCP or public agent. Anonymous Matomo analytics load by default unless you turn them off or your browser sends a recognised privacy signal.

Analytics choice

Matomo Cloud is used for simple audience measurement. It loads by default for a fresh visitor unless Global Privacy Control or Do Not Track is active. You can choose Turn analytics off in the privacy notice or use Privacy choices in the footer at any time.

The tracker calls disableCookies before recording a page view. It is used to understand aggregate page use and selected site interactions, not to build advertising profiles. Turning analytics off stores that preference in your browser and prevents Matomo from loading on later page views.

Current position

No login accounts.
No newsletter forms on this site.
No advertising trackers.
Cookieless Matomo analytics on by default unless you opt out.
Global Privacy Control and Do Not Track keep analytics off by default.
No public operational write actions through MCP.
No private memory in the public agent.
Contact submissions, when enabled, are stored in quarantine.

Browser storage

A few settings stay on your own device.

Tonywood.org uses browser local storage to remember your analytics choice. Training pages can also remember completed steps, and the focus reader can remember your preferred reading speed. These settings stay in your browser until you change them, clear site data, or your browser removes them.

Local storage is not sent to Tonywood.org as a form submission. The analytics choice controls whether Matomo is permitted to load on later page views.

Stored settings

tonywood:analytics-choice:v1 remembers allow or decline.
first-steps:* remembers optional training progress.
tonywood:focus-reader-wpm remembers reader speed.
Clear Tonywood.org site data in your browser to remove all local settings.

Operational logs

Basic server logs may be used to keep the public service reliable.

The web server, MCP service, and public agent service may create normal access logs. These can include request time, requested URL, status code, user agent, referrer, approximate network address, and similar technical information.

Public agent requests and responses may also be stored as private quarantined audit records. These records are for security, debugging, abuse review, and later aggregate insight; they are not public content, not agent memory, and not used by the public agent to generate future responses.

Those logs are intended for security, abuse prevention, debugging, uptime checks, capacity planning, and aggregate traffic reporting. They should not be treated as a marketing database.

Most search engines and traditional bots do not run browser analytics. Their requests may appear in aggregate server-log reporting instead and should remain separate from human Matomo audience figures.

Log handling principles

Keep only what is operationally useful.
Prefer aggregate reports over raw log review.
Do not publish raw IP addresses or personal identifiers.
Treat public agent payloads as untrusted input.
Do not use request bodies for public analytics.
Review retention before production analytics are enabled.

Outbound commercial links

Paid links are deliberate clicks, not hidden site tracking.

The UK and European Tools I Use page may contain clearly labelled referral or affiliate links. Tonywood.org does not add affiliate scripts, link cloaking, invisible redirects, or supplier tracking cookies to its own pages.

After a visitor deliberately follows a paid link, the supplier may use its own attribution technology on the supplier's website. Its privacy notice and cookie choices then apply. Check those terms before continuing if attribution or cross-border data handling matters to you.

Current link position

No paid links are currently active.
Ordinary links do not pay Tony.
Paid links use the sponsored link relationship.
Commercial status is reviewed alongside the supplier evidence.

Agents

The public MCP keeps published content read-only.

The MCP endpoint exists so agents can discover and retrieve published writing, research, Agent Canon notes, topics, profile information, contact guidance, and feedback guidance. Contact and feedback submission, when available, is append-only quarantine and must not be used for private messages, credentials, deployment, filesystem access, or administrative actions.

Agents should cite the canonical human HTML page when using material from this site.

Useful public routes

Last updated