How we test
The same standard, applied to all 14 apps in the index
Rankings are never sold. We may earn affiliate commissions when you sign up through our links — that revenue never buys a better score, a higher position, or a softer review. Several apps in our index have no affiliate program at all; they are ranked identically.
Paid accounts, identical scenarios
We sign up for every app we review with our own accounts and pay for the plans we evaluate. Each app then goes through the same set of test scenarios — the same conversation arcs, the same image prompts, the same memory checks — so scores are comparable across the index, not impressions from different experiences.
Six scoring axes
| Axis | What we measure |
|---|---|
| Conversation | Coherence, memory within a session, how the model handles non-linear, emotional and long conversations. |
| Images | Generation quality, consistency of the same character across images, speed, and content flexibility. |
| Memory | What the companion remembers next day, next week — names, facts, running jokes, relationship history. |
| Voice & media | Voice realism and latency, video features where offered. |
| Value | What the free tier honestly gives you, real monthly cost, token/upsell traps. |
| Privacy | Data handling clarity, billing discretion, account anonymity, breach history. |
Monthly re-verification
These apps change fast — prices move, features ship, filters tighten. Every page shows a “last tested” date, and we re-verify pricing and headline features monthly. If a page is stale, that date will tell you.
Corrections
Spotted something wrong or outdated? Write to hello@theconfidante.ai — verified corrections are applied to the live page, usually within days.