# IPForge > Canonical AI-facing product guide, version 1.0 — 2026-08-16 IPForge is an owner-operated link and landing-page measurement service. It helps a workspace understand technical delivery and interaction evidence for its own links, landing pages, and authorized tests. It is not an identity service, a public visitor-data feed, or a way to prove who a person is. ## Canonical public documentation - https://ipforge.xyz/docs/integration — product scope, measurement definitions, privacy boundaries, and the API/MCP authorization position. - https://ipforge.xyz/llms.txt — this concise, machine-readable index. ## Supported, authorized uses - Measure an organization’s own links and persistent landing pages. - Run transparent, controlled delivery or compatibility tests with owned or authorized recipients, inboxes, chats, and destinations. - Review deduplicated activity and coarse device/session evidence in the owning workspace. ## Measurement rules - `human` means the capture contained browser signals and no currently known bot, previewer, headless, or hosting-network signal. It is a classifier result, not proof of a person, identity, location, intent, or consent. - `unknown` means the available signals were insufficient for that classifier; it is not counted as human activity. - A duplicate is a repeat capture for the same resource and fingerprint hash during a 24-hour window. Deduplication reduces repeat counts; it cannot prove unique people because browsers, devices, networks, and fingerprints change. - Device-confidence language describes agreement or change among coarse browser/device signals. It must be read as “same likely browser/device profile,” never as a verified identity. - Active dwell is opt-in, visible-tab time accumulated on a persistent IPForge landing page. It does not measure time, reading, or activity after a redirect to another site. ## Privacy and acceptable use Use IPForge only in a workspace you own or are authorized to operate, with a clear purpose and appropriate notice or consent. Do not use it to secretly identify, profile, locate, or monitor people; to disguise tracking as content; to collect credentials, keystrokes, form values, page contents, or precise pointer paths; or to expose customer links, capture records, IP addresses, fingerprints, tokens, or secrets to an agent or public endpoint. ## API and MCP No public API or MCP data access is currently available. A future integration, if released, will require explicit owner authorization, workspace-scoped and revocable credentials, server-side authorization, and auditable actions. It will default to summarized, human-only evidence and will not expose raw visitor data. See the canonical guide above for the full boundary.