Making Voice More Affordable and Scalable

After sustained optimisation, ChatMoni’s in-house voice pipeline reached a key milestone: lower internal cost per minute and a capacity model tied to infrastructure we control.

Optimisation changed the equation

Building a voice engine was only part of the challenge. To make it useful for businesses, we also needed to improve its operating economics. By March 2026, continued optimisation had reduced the pipeline’s cost per minute below the comparison point we were targeting, even after accounting for our intended margin.

Capacity we can plan around

An in-house system changes how capacity can be expanded. Rather than being entirely dependent on a vendor’s concurrency tier, capacity can be planned around the GPU resources we deploy, the performance we achieve, and the workload we need to support. That does not make capacity infinite. It makes the scaling model more directly connected to infrastructure we can measure and extend.

Voice personalisation on our infrastructure

The pipeline was also intended to support users training and running their own cloned voices on ChatMoni infrastructure. Personalised voice can make an assistant feel more aligned with a business, but the experience must be supported by clear consent, access controls, and protections against impersonation or unauthorised use.

Lower cost is a product feature

Voice becomes much more practical when the cost structure can support regular use. Lower operating costs can create room to offer a useful experience at a price that small businesses can realistically consider. The goal is not merely to win a benchmark; it is to make voice commercially viable.

A milestone, not the end of the work

The March milestone represented progress in efficiency and control. Ongoing performance monitoring, infrastructure planning, reliability work, and responsible voice-use safeguards remain essential as usage grows.