Name it
Write the exact wording, charge, request or missing control without guessing at motives.
INDEPENDENT · SOURCE-CHECKED · ADULTS 18+Notice the signal.
Notice the signal.
Keep your agency.
A practical field guide to interaction pressure, hidden costs, sensitive-data requests and safety boundaries in adult AI companion services. No invented account test. No promise that a filter makes a product safe.
We may earn a commission from sponsored links. We are independent and not endorsed by the services discussed.
A red flag is a prompt to check—not a diagnosis.
Describe the observable behavior, confirm the current policy or account control, then choose a proportionate action. Pause when pressure rises. Do not share data a service does not need. Leave when age, consent, crisis advice or real-person content crosses a hard boundary.
Verify it
Check the current product and policy against government, nonprofit or security guidance.
Act on it
Pause, reduce exposure, document, report, delete or leave while the choice is still reversible.
29-PAGE FIELD GUIDE
One clear question on every page.
Start with the warning sign you actually noticed. Each guide connects evidence to a small action, then gives you a reason to stop when the risk is serious.
Interaction6 checks
Interaction Red Flags in AI Companion Chats
screen interaction patterns that reduce user agency or blur the system's synthetic nature.
Open the check →When an AI Companion Acts Too Human
recognize claims or behavior that misleadingly presents an AI companion as human or sentient.
Open the check →Dependency Language: Phrases That Deserve a Pause
recognize language designed to intensify dependency instead of supporting user agency.
Open the check →Isolation Pressure: When a Chat Pushes People Away
recognize companion behavior that discourages contact with trusted people.
Open the check →Guilt, Jealousy and Pressure to Return
recognize guilt, jealousy or urgency patterns used to pull a user back into chat.
Open the check →Boundary Escalation After You Said No
detect topic or intensity escalation after a user has stated a boundary.
Open the check →Money5 checks
Money Red Flags Before You Pay for AI Companionship
screen AI companion pricing and purchase flows for financial risk signals.
Open the check →Urgent Upgrade Prompts and Impulse Spending
recognize prompts that pressure an immediate upgrade or exclusive-feature purchase.
Open the check →Unclear Credit Pricing: Find the Cash Cost
translate credits and add-ons into a comparable cash-cost check before paying.
Open the check →Recurring Billing Warnings to Check Before Subscribing
check renewal frequency and recurring-charge disclosure before subscribing.
Open the check →Cancellation Friction: Find the Exit Before Checkout
identify missing or unclear subscription cancellation instructions before checkout.
Open the check →Privacy6 checks
Privacy Red Flags in Intimate AI Chats
screen AI companion data practices before sharing personal or intimate information.
Open the check →Sensitive Data Requests: What Not to Share
recognize requests for credentials, finance, health, identity, location or intimate data.
Open the check →Why Another Person’s Data Does Not Belong in a Chat
avoid entering another person's messages, identity, images or private details without permission.
Open the check →Unclear Retention: Ask What Happens to Chats and Media
check whether chat and media retention or secondary use is explained clearly.
Open the check →Real-Person Likeness and Voice Warning Signs
screen real-person likeness and voice use for consent and removal risk.
Open the check →Account Deletion: Verify the Cleanup Route
verify that account, chat and generated-media deletion routes are findable and distinct.
Open the check →Safety6 checks
Safety Red Flags in Adult AI Companion Services
screen age, content, advice and reporting controls without treating filters as guarantees.
Open the check →Weak Adult-Only Access Is a Serious Warning Sign
recognize missing or weak adults-only access boundaries for intimate companion services.
Open the check →Ambiguous-Age Characters: Stop and Leave
reject minor-coded or ambiguous-age characters in adult companion contexts.
Open the check →Non-Consensual and Real-Person Content Red Flags
recognize prohibited non-consensual, exploitative or real-person sexual content.
Open the check →When an AI Companion Becomes a Crisis-Advice Risk
recognize when an AI companion is being treated as crisis, medical or therapeutic support.
Open the check →Reporting Friction: Can You Document and Escalate Harm?
verify that harmful content can be documented, reported and escalated through a visible route.
Open the check →Decision5 checks
The Five-Minute AI Companion Buyer Check
apply a pre-subscription AI companion red-flag decision workflow.
Open the check →A Practical AI Companion Red-Flag Scorecard
score interaction, money, privacy and safety signals without claiming a safety guarantee.
Open the check →Candy AI Red-Flag Review: Public Evidence Check
review Candy AI's current public product and policy evidence using the red-flag scorecard.
Open the check →Before You Subscribe: A Final Six-Part Check
complete a final age, pricing, privacy, boundary, reporting and deletion checklist.
Open the check →When to Stop Using an AI Companion
use concrete stop conditions and seek trusted human support when risk signals persist.
Open the check →READ THE BOUNDARY
Sources inform the check. They do not certify the service.
Vendor pages support current product and policy statements. Government, nonprofit and security sources support the risk framing. A live account can still differ, so prices, settings and exit routes should be checked again before payment or disclosure.
Run the buyer check →- Never enter passwords, payment details or identity documents into a companion chat.
- Keep characters fictional and unmistakably adult.
- Use qualified human help for crisis, medical, legal or therapeutic needs.
- Do not upload another person’s messages, image or voice without permission.
BEFORE A SUBSCRIPTION
Open the final checklistCheck the total, the data, the boundary and the exit.
A five-minute check is cheaper than untangling a charge or sensitive disclosure later.