How I priced my AI SaaS at $6.90/mo — and why I charge differently in Brazil
I've been going back and forth on pricing for months. TubeSpark is an AI co-pilot for YouTube creators, and I finally landed on $6.90/mo for Starter and $17/mo for Pro. Getting there was not clean.
The $6.90 problem
Most AI SaaS tools charge $19-29/mo minimum. I almost did the same. But my target users aren't enterprise teams or funded startups. They're YouTube creators in Brazil, Mexico, India, and Southeast Asia who make content as a side hustle. $29/mo is their entire monthly entertainment budget.
I looked at what a mid-tier YouTube creator in Brazil spends on tools. Canva Pro is R$35/mo. Epidemic Sound is R$22/mo. Their total tool budget is maybe R$150/mo. If I charged $19 USD, that's R$110 at current rates, which is 73% of their budget for ONE tool. No way that works.
So I went low. $6.90/mo USD for international. R$47/mo in BRL through a separate Brazilian payment processor.
Why two payment systems
Stripe doesn't do PIX (Brazil's instant payment system that literally everyone uses). Most Brazilian creators don't have international credit cards either.
I ended up running Stripe for USD and LemonSqueezy for BRL. Two webhook handlers, two subscription flows, two sets of edge cases to debug at 2 am. Not elegant. Works though.
The math
At $6.90/mo, my margins are thin. I route through 4 AI providers (GPT-4, Claude, Gemini, Groq) and the API costs depend on which models each user triggers. A Pro user generating 45 scripts/month costs me $2-3
in API calls. Starter users mostly generate ideas which hit the cheaper/faster models.
Starter users stay in the idea zone, and only Pro users use the expensive script pipeline. If that's wrong I'll need to rethink the tiers. I don't have enough data yet to know.
Mistakes I already made
Should have launched with annual pricing. Added it later, and the conversion rate for annual is still embarrassingly low because nobody trusts a new product enough to pay for a year upfront. Also validated the BRL price by asking 3 friends. One of them doesn't even use YouTube. Great research methodology.
The actual question
I don't know if $6.90 is too low. Some founders here have told me I'm leaving money on the table. Others say low price + high volume is the play for emerging markets.
How do you price for markets with different purchasing power? Start low and raise, or start high with regional discounts?
The two-payment-system situation is real pain — but you've already solved the hardest part (actually shipping it).
On pricing: 'start low and raise' works better than 'start high with discounts' in emerging markets, for exactly the reason you described — trust. Nobody pays for a year upfront on a new tool. But .90 × 12 = 2.80 annual, and after 6 months of actual use they'll pay 0 annual no problem.
One thing worth thinking about at .90/mo: failed payment recovery matters more at low price points, not less. If a Brazilian user's international card declines (which happens constantly with cross-border transactions), they probably won't bother to update their card for .90 — not because they don't value the product, but because the friction of fixing it exceeds the perceived urgency. Your Stripe churn is likely silent and you won't see it until renewal time.
The LemonSqueezy + PIX path probably solves this better for BRL users since PIX is instant and rarely fails. But for your USD Stripe subscribers, a good dunning sequence on first payment failure will quietly save 20-30% of those churns.