Introducing Growth & Scale: guarantees that cost more, on purpose
Two new tiers sit above self-serve - not because we discounted anything, but because uptime, reserved throughput, and real support response times are worth paying for honestly.
Self-serve has always been simple: pay as you go, no promises attached. That works well for a side project or an early prototype. It stops working the moment your product is in front of real users and a slow response, a rate limit, or an unexplained gap in support becomes something your customers feel.
That's the gap Growth and Scale are built for - not enterprise, not another self-serve discount, but an honest middle tier for teams that have outgrown "no guarantees" and don't need a procurement conversation to get what they actually need.
What you're actually paying for
Neither tier is ever cheaper per token than self-serve. That's deliberate. You're not paying less for guarantees - you're paying a fair premium for them, the same way reserved-capacity tiers work at any serious inference provider. What changes is what you get on top of the same underlying models:
- Growth - KES 6,500/month + KES 92 per 1M tokens. 99.9% uptime SLA, 500 requests/minute reserved, 48-hour support response, public status page access.
- Scale - KES 18,000/month + KES 72 per 1M tokens. 99.95% uptime SLA, 2,000 requests/minute reserved, 24-hour support with a shared Slack channel, monthly usage reporting.
Scale's per-token rate is lower than Growth's - not because it's a discount, but because a bigger, more predictable customer costs us less to serve reliably per token than a smaller, more variable one. The platform fee is what funds the guarantee itself; the usage rate is what funds keeping the guarantee true every day.
Billing that doesn't surprise you
Two separate charges, on two separate rhythms. The platform fee is fixed and charged upfront on your billing date - no ambiguity about what that number covers. Usage draws down from a prepaid balance in real time, the same instant a request completes, with auto-recharge before you ever hit zero. Nothing bills in arrears, and nothing about your bill should ever be a surprise a month later.
If you already know it's a big number
If your team is planning around hundreds of millions or billions of tokens a month, Growth and Scale aren't the right shape for that conversation - a reserved-capacity quote is. Tell us the number and we'll talk through a committed-volume rate directly, without pushing you through a tier built for a smaller footprint.
Growth and Scale are live today, no long-term contract required on either - switch tiers or cancel anytime.
James Miano
CTO & ML Engineer at Roniki Systems. James specializes in low-overhead LLM quantization processes, custom ternary weights architectures, and localized server optimization.
Stop paying for overpriced round-trip latency
Why route queries over Western servers when you can use low-overhead hardware located in Nairobi? Save 87% on your monthly inference spend. No minimum credit limits. M-Pesa ready.