The first 100 users get 500 free credits. Join waitlist

Wolfvisibility

Documentation

Guides for setting up prompt sets, running checks, and verifying evidence. Full docs are expanding ahead of launch.

Core workflow

From buyer question to auditable observation

Wolf visibility is built around defined checks rather than a shared black-box index. One check combines a prompt, an engine, and a regional or language target.

Build a useful prompt cohort

Start with real buyer questions: category discovery, comparisons, alternatives, problems, product requirements, pricing, and regional intent. Keep benchmark prompts stable when measuring change.

Choose engines and targets

Select the AI surfaces and supported regional or language targets that match the market. Wolf previews the required credits before execution.

Read the evidence

Review the raw answer together with detected brands, relative position, citations, and exact source URLs. A score without its evidence can hide important context.

Compare responsibly

Compare like with like: the same prompts, compatible engines, the same target scope, and successful responses. Record changes without treating correlation as causation.

Important operating notes

AI answers can vary between runs because engines change their models, retrieval systems, available sources, and product behavior. Location, language, time, and account context may also influence a response. Wolf preserves the metadata and raw evidence for a defined observation; it does not claim that one response is a fixed rank or represents every user.

A failed engine execution is not the same as a brand omission. Failed checks are recorded separately, excluded from successful-response metrics, and not charged. When an engine exposes citations or grounding sources, Wolf stores those URLs; when it does not, Wolf does not fabricate a citation layer.

Read the full methodology for metric definitions, or use the pricing calculator to estimate a prompt cohort before joining the launch waitlist.