Launch
Lighter, not bigger.
A lifelike AI avatar, drawn on the visitor’s own phone or laptop. One photo to create it, and no GPU on our side to draw it. How we got there, and what it lets you build.
Read article
Toda Labs
Research & updates
Engineering notes, product updates and research from Toda Labs.
More articles
Off the Grid · Part 4
Why a live AI avatar keeps a GPU busy for the whole conversation
An AI writes a reply in a second or two, and the GPU moves on. A live face rendered in the cloud keeps a data-center GPU working for the whole conversation, even while it only listens.
Read article
Off the Grid · Part 3
8 questions to ask an AI avatar provider about pricing and capacity
Eight questions about capacity to put to any AI avatar provider before you sign, from what each conversation holds to where the GPUs come from, with our own answers.
Read article
Off the Grid · Part 2
Why live avatars are billed by the minute, and what that adds up to
A cloud rendered realtime live avatar’s GPU time doesn’t stay in the data center. The business and end users pay for it, minute by minute. Let's look at what a month of those minutes would cost, for two examples: a support desk whose avatar talks with customers for 500 hours, and a training app with 600 users who practice every week.
Read article
Off the Grid · Part 1
A live avatar that doesn’t hold a data-center GPU for the conversation
A live avatar asks a data centre for a GPU for every second of a conversation. What that holds in power, water and grid capacity, and what changes when the face renders on the device.
Read article
Engineering note
Why MeshLite animates avatars in the browser
Why we chose browser animation, and what that means for longer conversations, repeated use and avatar-rendering costs.
Read article