Free AI Story Companion: $6-13 Pass, Microtransactions, Retention Checklist
A nights-and-weekends builder can pair free user stories with a $6-13 pass, tiny purchases, and a retention check.
You have a free AI story companion that people finish. Zeta AI Companion earned more than $10 million in monthly revenue in August 2026. The question is where your first dollar lands without scaring the people who made it fun.
A companion app can earn without a content team if the free layer supplies the story, the paid layer removes friction, and the tiny purchases keep the habit going. The first dollar should land on friction, not on the story itself.
Zeta's Japan first-quarter revenue grew 691% year over year, and its average revenue per user was above $10. The scale does not require a team; it shows the shape of the stack.
The free layer supplies the story
Character.AI provides millions of user-created characters, a free tier, and no card-collecting mechanic. That model keeps the story supply alive because the people making it are not being squeezed out. A community supply chain gives you more story arcs than a weekend builder can write alone.
Zeta does not pay its user-generated content creators for the storylines they build. If your weekend app depends on UGC storylines, a missing creator payout is a fixable issue before it becomes a churn problem.
Talkie AI's free plan allows 50 messages per day and 2-minute voice calls, while its paid plan costs $9.99 monthly. The free tier can be generous and still leave room for a paid pass, with limits that protect the pass rather than punish the reader.
Character.AI's 2026 free tier included full-screen advertisements during conversations, a practice that annoyed some users. Ads can fund a free tier, but they can also make the free layer feel like a waiting room.
The checklist is a Sunday test
The Hook-and-Hold Weekend Stack is the part you can test in a short sitting: a free UGC hook, a $6-13 subscription, microtransaction props, a retention metric, and a creator-payout floor. Each item is a yes/no check, not a growth theory. The early items protect the free layer; the later items keep the paid side honest.
- Confirm the free layer lets a reader finish a story arc.
- Set the pass price in the Zeta Pass range.
- Keep basic conversation open, not locked behind a collecting mechanic.
- Add a small purchase for a story add-on.
- Make the paid pass remove a named friction, not the whole story.
- Put a retention metric in a dashboard.
- Write a creator payout floor before paid packs ship.
Crushon AI and Candy AI use credit systems for selected features, while basic conversation is not locked behind a collecting mechanic. That keeps the core habit free.
Zeta sells story packs for ¥300 and character outfits for ¥100 using its Pieces virtual currency. The tiny price serves the user who is not ready for a subscription.
Tendera charges $9.99 monthly for premium access and avoids gem purchases, card pulls, and token counters. The clean premium plan still feels optional.
The math keeps the weekend honest
Paid plans usually reach less than 5% of users, with the rest using the service without upgrading. Imprezia's AI companion monetization math puts inference costs 10 to 30 times higher than traditional apps and subscription conversion around 3%. The pass must be cheap enough to clear that gap. The conversion number is the reason the pass has to feel like a tip.
Zeta's usage time among South Korean smartphone users in February was 113.41 million hours, more than twice ChatGPT's 50.47 million hours in the same period. Usage time shows whether the companion is becoming a habit. For a weekend builder, that signal is the cheapest proof that the story is working.
Most side projects stay side projects, and that is fine; the stack is small enough to run on weekends and clear enough to stop when the retention metric stops moving.