The question we've been asked more than almost anything else: do I really have to record hours of video to build a custom AI avatar?

The answer is no. One image is enough.

We just launched our ONE-SHOT feature. This technology allows users to create a custom avatar with a single image. Our models then takes this image and generates a fully conversational interactive AI avatar from it. No studio session. No lengthy onboarding. Upload an image, and within a couple of minutes you have an avatar that you can interact with in real-time.

How does ONE-SHOT actually work?

The interactive avatar generated from your image isn't pre-rendered animation. It's a live model running on our image generation models architecture, which generates every pixel in real time from audio input. When someone talks to a custom avatar, the avatar responds the same way it would if you'd recorded hours of training data. The slight head movement when listening, the blink timing, the micro-expressions that make the difference between "obviously AI" and "wait, is that real?"

That's the problem we've spent 18 months on. Not making AI avatars that look good in a demo. Making interactive avatars that hold up across a full-length conversation with someone who's actively looking for the seam.

The difference shows up most in the details. The way our avatars pause before answering a difficult question. The way it tracks when someone interrupts. These aren't scripted behaviours.

Who's building with this?

More than 8,000 organisations are already using Anam. The use cases have expanded faster than we predicted.

Replicate Labs scaled their sales coaching operation using AI interactive avatars. Sales reps practice with an avatar that plays the customer, handles objections, and scores the conversation. CoachHub is delivering personalised coaching at a scale that wouldn't be possible with human coaches alone. Preply offers native-language conversation practice on demand, at any hour, in any language we support.

These aren't experimental deployments. They're production systems running at scale, and the AI talking head technology behind is Anam's real time avatar models.

The use cases developers bring to us are often ones we didn't anticipate. Interactive AI avatars can be user as the default interface across a number of use cases.

What the funding makes possible

We raised $9 million from Redpoint, Torch and SV Angel. That funding is going directly into the model work, making our real time models faster and more expressive at lower cost.

Custom avatar creation is currently available to all plans.

Developers: the API is free to try

If you want to build with this, the fastest path is a free account on Anam Lab. You get API keys, access to the SDK, and documentation without a sales call or a credit card. The ElevenLabs integration is also live if you're already running a voice agent and want to add a face.

If you're thinking about what digital humans can do for your product specifically, the use case thinking is worth reading before you start building.

Frequently asked questions

What is ONE-SHOT technology?

ONE-SHOT is a feature that enables Anam's users to create custom avatars with a single image.

Is custom avatar creation available to all customers?

Yes.

What makes Anam's AI avatars feel human?

Anam's state-of-the-art real time video generation models generate avatars who feel as natural and engaging as interacting with a human live on a video call.

How do I get started building with Anam?

Create a free account at Anam Lab. You get API keys and SDK access with no credit card required. Documentation and integration guides are available immediately after signup.