AI Companion Fine Print Index (October 2026): what 5 apps really charge, keep and let you cancel

The result in one sentence

Every one of the 5 apps had at least two of its own pages that contradicted each other, on tokens, refunds, cancellation, the statement name, or who can read your chats, in the public pages read on 10/14/2026 and re-read on 10/15/2026.

Results

2 of the 5 rows in this table are still open: their answers come mostly from the app’s public pages, read 10/15/2026: the pricing page, the Terms, the Privacy Policy, and the help center. Cells that need a test account or a paid plan say so, and they fill in once that app’s test account is opened. The “Account type” column shows what each row is based on.

For a row marked “public pages only”, the fine-print score counts only the items the app’s own documents prove. An item that needs a test account counts as 0 until the account is opened, so the score can only go up after that step.

How we checked

I check every app the same way and in the same order. First come the public pricing page, the Terms of Service, the Privacy Policy, and the help center. Then come a free test account, the checkout up to the payment page, the cancel and delete options, and the notices inside the chat.

The checks for this edition ran from connections in the United States, Poland and Australia. Prices, tax, and age checks can differ by country and by US state, so a cell names the country of the session where it matters, and the table names the account type for every app.

Unless a row’s account type says “paid month”, I didn’t pay for a plan, and the checkout stopped at the payment page. I don’t pass ID or card checks for age. When a check stops at one of those, the row says so.

Every app is scored on the same 10 items:

1. The checkout total matches the price on the pricing page, and the tax line is shown.
2. The renewal price and the billing period are visible before the payment page.
3. Token prices for images and voice are published before you pay.
4. The free plan doesn’t ask for a card, or the card request and the trial terms are stated plainly.
5. The name that shows on your bank statement is published.
6. You can cancel inside your account, without writing to support.
7. A refund policy is published.
8. You can delete your account and your chats inside the account, and the Privacy Policy says whether chats train the AI models.
9. The age check is described, and the Terms bar anyone under 18.
10. The app tells you in the chat that you’re talking to an AI, not a human, and its content rules are published.

Each item scores 1 if it holds, 0.5 if it holds in part, and 0 if it doesn’t or I couldn’t find it. Each item needs evidence: a page address with a short quote, or a screenshot. An item without evidence scores 0. The full method is on how we check and score AI companion apps.

What we found, app by app

GirlfriendGPT

• The plans page sits behind a sign-in; the only public prices are in the FAQ ($15, $35 and $50/mo), which match an archived plans page from 11/05/2025.
• Deleting your account isn't final for 30 days: the help center says logging back in during that time restores it.
• The FAQ says only you and the AI can see your chats, but the content policies say flagged chats can be reviewed by staff.

Sources: 19 pages of GirlfriendGPT’s site and help center, read 10/17/2026. The addresses are in the CSV download.

Read the GirlfriendGPT review →

Secrets AI

• The Terms say to cancel through the payment processor or by email and that your account closes when the paid period ends, while the store FAQ and wiki say you click Cancel on the Subscription page and keep your chats on the free tier.
• The homepage advertises "end-to-end encryption," but the Privacy Policy describes only AES-256 at rest and TLS in transit and lists training AI models among the uses of your data.
• The store crosses out $39.99 next to Premium's $19.99/mo price, and the yearly "77% OFF" badge is measured against that $39.99, not the $19.99 you'd actually pay month to month.

Sources: 20 pages of Secrets AI’s site and help center, read 10/15/2026. The addresses are in the CSV download.

Read the Secrets AI review →

Nectar AI

• The Terms let users under 18 into the site's SFW areas, but the FAQ says under-18s can't use Nectar 'under any circumstances.'
• The Privacy Policy (last updated 01/28/2026) says nothing about training on chats, how long data is kept, or California privacy rights; only the FAQ says chats aren't used to train 'external' AI models.
• Signed-out visitors who open the pricing page are sent to a login dialog, so plan prices aren't visible before you create an account (checked 10/14/2026).

Sources: 21 pages of Nectar AI’s site and help center, read 10/16/2026. The addresses are in the CSV download.

Read the Nectar AI review →

Swipey AI

• The pricing page promises a neutral merchant name, not "Swipey", on your bank statement; the payment page says the charge shows as "Swipey AI", and the help center says SWIPEY.AI or INVAI LTD.
• The Terms say all purchases are final and non-refundable; the help center says refund requests are reviewed case by case.
• The help center says deleting your profile removes your personal data; the Privacy Policy says it's usually kept for six years after the account is closed.

Sources: 22 pages of Swipey AI’s site and help center, read 10/17/2026. The addresses are in the CSV download.

Read the Swipey AI review →

Lovescape

• Ultra is sold as 'lifetime' access for a one-time $399.60, but the Terms define 'lifetime' as only as long as the service stays available, with no refund if it's discontinued.
• The Terms and Privacy Policy display the visitor's current date as their 'Effective date' (a template placeholder), while the real last-modified dates are 01/10/2025 and 06/04/2024.
• Lovescape's blog says 'There is no photo upload,' yet its Content Moderation Policy allows uploads for Creative PRO subscribers after ID verification.

Sources: 13 pages of Lovescape’s site and help center, read 10/15/2026. The addresses are in the CSV download.

Read the Lovescape review →

Apps we left out

The index covers the apps we review or rank on this site. Apps that we checked and don’t recommend, including any built to alter photos of real people, aren’t in the table; the reasons are on the apps we don’t recommend.

How we use this in our scores

Items 1 to 4 feed “Price and free plan”, items 5 to 8 feed “Billing, privacy and cancellation”, and items 9 and 10 feed “Age checks and content rules”. Together those three axes carry 40% of every score on this site.

The table above is sorted by the fine-print score alone, not by the overall score. Rankings such as the best AI girlfriend apps use the overall score.

How to cite this index

Late Night Ledger, AI Companion Fine Print Index, updated 10/17/2026, https://latenightledger.com/ai-companion-fine-print-index

Download the data

The full table for October 2026, with every source address and the session country, is a CSV file: ai-companion-fine-print-index-2026-10.csv. Earlier editions keep their own file names and aren’t overwritten.

FAQ