Don't guess if an AI companion is legit.

Your AI companion evidence check

Open the service's terms, privacy notice, help pages and final checkout. Answer yes only when you can point to clear current evidence.

What “legit” should mean before you pay

A real website can accept payment and still be a poor choice. A popular service can have clear policies and still not fit your needs. Instead of asking for one magical yes-or-no label, break legitimacy into evidence you can check: who operates the product, what you are buying, how renewal works, how private content is handled, whether support responds, and how you can leave.

The Trust Score turns those questions into a consistent pre-payment routine. You answer yes only when you found clear evidence. Choose unknown when the answer is missing or contradictory. Choose no when the service clearly lacks the safeguard described. The result is an evidence-gap score, not a claim that a company is good, bad, safe or fraudulent.

Why an interactive check beats a borrowed verdict

Reviews are useful for discovering features and recurring complaints, but they are snapshots. Prices change. Help pages change. A cancellation route that worked for one person may differ when you subscribe through an app store. Your region, device and chosen plan can also change what appears at checkout.

A repeatable check forces the current first-party evidence into the decision. It also makes comparisons fair: every option faces the same questions. If one service has impressive media examples but unclear renewal terms, that uncertainty remains visible instead of disappearing behind an overall star rating.

The five evidence folders

Start with company identity. Connect the public brand to the operator named in the terms and privacy notice. Then review money: final amount, currency, interval, credits and renewal. Next examine private content: chats, uploads, generated media, retention and deletion. Confirm how you can leave by checking both cancellation and account deletion. Finally, test the support route with one simple question before the stakes are high.

You do not need a legal background. Save the pages you relied on, date your notes and avoid guessing. A missing answer can often be resolved with one support message. If it cannot, the score should reflect that uncertainty.

How to use the result

A lower evidence-gap result means you found more clear answers. It is not permission to share sensitive material or spend more than you planned. A middle result usually means several important questions remain unresolved; complete the suggested actions and score again. A high result means too many basics are missing or negative for a comfortable payment decision.

The most useful part is the action list below the number. It identifies the questions that contributed most to the result. Work through those items in order. If the service provides convincing current evidence, change the answer. If not, postpone payment or compare another option with the same standard.

A safer first-session routine

Use a separate email, keep the first conversation fictional, avoid personal photos and do not reveal a full name, workplace, address or identifying story. Test the free level if one exists without a card. If you pay, choose the smallest commitment that can evaluate your main use case and save the final checkout screen.

Set your own renewal reminder immediately. Confirm where cancellation lives while the account is new. These habits are useful even when a service has clear policies because mistakes and changing priorities do not require bad intent.

What this score deliberately does not do

The tool does not scan a website, access private systems or certify compliance. It does not assign secret ratings to brands. It does not know whether a policy statement is technically enforced. It does not replace professional legal, financial, security or mental-health advice.

Its job is narrower and practical: stop an impulsive checkout long enough to gather the evidence a reasonable buyer would want. That makes the score useful without pretending it can eliminate every risk.

A worked example without a brand verdict

Imagine a service whose terms identify an operator, whose final checkout clearly shows a monthly renewal, and whose help page explains cancellation. Those answers can be marked yes. Now imagine that its privacy notice never explains whether chat text is used for product improvement, the media deletion route is missing, and the support address has not replied. Mark those items unknown rather than deciding the service must be safe or unsafe.

The result will rise because important evidence is unresolved. The action list should point first to private-content use and deletion, not to a cosmetic detail. You can send two precise support questions and wait. If clear answers arrive and match the published policy, update the inputs. If answers remain absent, the number has done its job: it made the uncertainty visible before payment.

Now run the same process on a second option. Do not reward it merely for having a lower headline price or more dramatic examples. Compare the evidence for the use case you actually want, the complete likely cost, and the exit path. The tool helps you make that comparison consistently, but the final boundary is yours.

Investigate one question at a time

Ready to compare current options?

Named services are optional starting points, not outcomes of the score. Apply the same evidence standard to each one.