minai is a pay-as-you-go AI assistant built for virtual assistants in emerging economies — starting in Kenya, Nigeria, and the Philippines — where frontier AI tools like ChatGPT and Claude are inaccessible due to credit card requirements and $20/month price tags.
By integrating with MiniPay's stablecoin wallet, minai enables users to top up their AI credits with as little as $0.10 in cUSD, USDC, or USDT on the Celo blockchain. There are no subscriptions, no bank accounts or debit cards required, and no wasted spend — users only pay for the tokens they consume.
AI shouldn't be a luxury.
The global remote workforce — virtual assistants, freelancers, and knowledge workers in emerging markets — represents millions of people whose livelihoods depend on staying competitive. As AI reshapes how work gets done, access to these tools is no longer optional. It's existential.
minai wants to close that gap. By pairing frontier-grade AI with Celo's stablecoin infrastructure, we make it possible for anyone with a smartphone and a dollar to work smarter, deliver more, and protect their income. Not someday. Today.
Because the workers who have the most to gain from AI shouldn't be the last ones to get it.
Virtual assistants in emerging markets are caught in a paradox: the same AI tools threatening to automate their jobs are also the ones they can't afford to use defensively. A VA in Lagos or Manila netting $3.50/hour cannot justify a $20 to $100 per month ChatGPT subscription. And even if they could, the likelihood that they have a credit card is under 1 in 20.
Without access to AI, they fall behind clients and competitors who use it daily — slower turnaround, more errors, less output. The irony is that VAs stand to benefit more from AI augmentation than almost any other knowledge worker: their job is inherently task-oriented, repetitive, and time-sensitive.
The tools exist, but the access doesn't.
Under the hood, minai uses multiple optimizations to reduce cost.
First: we wrap Qwen 3.5 Plus on the Alibaba Cloud, which is a frontier-grade multimodal LLM that's 80% cheaper than Claude Opus.
Next, we analyze each user prompt by first running it through a fast classifier which then routes it to the most cost-effective model available, which avoids slow response times and reduces costs even further. (Not every query needs deep reasoning, i.e., the most expensive model.)
Finally, we do background compacting that reduces the input context size by as much as 90%, while still maintaining overall accuracy.
The result is a user experience that is 90% as performant as Western frontier-grade models, but delivered at just 10% of the cost. And since it's all pay-as-you-go, the impact on the customer is extremely lightweight.
minai is specifically tailored to virtual assistants, with built-in tools for Google/MS Teams Calendar management, client notebook organization, document analysis, and headless browsing — the everyday tasks that define their work.
We're also working on a personal OpenClaw enclosure that will allow them to design their own skills for tasks that we haven't considered.
Lastly, there's a bounty program that allows them to earn more LLM tokens purely by making accepted feature suggestions.
minai demonstrates how stablecoin micropayments on Celo can unlock access to transformative technology for the workers who need it most.
Pay-as-you-go transactions
Want to raise from VCs
Initially Nigeria, Kenya, Philippines
