Privacy model

We measure audiences, not people.

This page is the plain-language version of the machine-readable manifest every Vector 5 site publishes. If the two ever disagree, the manifest is wrong and we want to know.

No cookies No localStorage IDs No fingerprints No stored IPs Daily-rotating hash Aggregates-only AI

What we collect

Every field, and why it exists.

Field	Why	Stored?
Page path	Traffic reports	Yes
Referrer host / source	Attribution	Yes — host only, never the full URL
UTM parameters	Campaigns	Yes
Country / region	Geo breakdown	Yes — IP discarded after lookup
Device, browser, OS, screen width	Device reports	Yes
Event name and optional props	Goals	Yes
Actor and AI source	AI traffic	Yes
Visitor hash	Unique visitors for one UTC day	Yes — unrecoverable, rotates daily
IP address	Hash + geo only	Never
Cookies / localStorage IDs	—	Never
Fingerprints	—	Never

How unique visitors work

A hash that forgets you at midnight.

The salt rotates at 00:00 UTC per site. After hashing, the IP is dropped. A visitor who returns tomorrow is counted as new. That is intentional: it gives accurate daily uniques while making cross-day tracking impossible, even for us.

visitor hash

visitor_hash = SHA-256(daily_salt || site_id || ip || user_agent)

daily_salt   rotates 00:00 UTC, per site, never logged
ip           used once for the hash and a country lookup, then discarded
user_agent   used for the hash and device/browser/OS parsing

Actors

Every event is tagged.

human

Browser visitor

human_via_ai

Human who arrived from an AI assistant

ai_agent

Headless agent acting for a person

ai_crawler

GPTBot, ClaudeBot, PerplexityBot and peers

bot

Other automated traffic

Crawler hits come from the server middleware. They never require JavaScript and never carry a visitor hash for a person.

What we cannot know

Who a visitor is
Whether they returned after midnight UTC
Cross-site or cross-device identity
Exact address (city is optional and off by default)
Anything about visitors who block the request entirely

The same list is rendered inside every dashboard, next to the numbers, so nobody forgets what the numbers are not.

Collection manifest · GET /api/sites/{domain}/manifest

GET /api/sites/example.com/manifest

{
  "name": "Vector 5 Analytics",
  "site": "example.com",
  "cookies": false,
  "local_storage": false,
  "fingerprinting": false,
  "ip_stored": false,
  "visitor_hash": "sha256(daily_salt || site_id || ip || user_agent); salt rotates 00:00 UTC",
  "retention_days": 730,
  "fields": ["path", "referrer_host", "utm", "country", "device",
             "browser", "os", "actor", "ai_source", "event_name"],
  "cannot_know": [
    "Who a visitor is",
    "Whether they returned after midnight UTC",
    "Cross-site or cross-device identity",
    "Exact address or city",
    "Anything about visitors who block the request"
  ]
}

AI insights

When enabled, insights are generated from rollups — counts, ratios, deltas. Event-level data is never sent to a model. With Ollama, nothing leaves your server at all.

Source: privacy-model.md Read the docs