Add a face to your
Elevenlabs Agent

Add a face to
Elevenlabs Agents

before

before

[0]

ElevenAgent Orb

Phone call

Get started in minutes not weeks

Server Side

Client Side

1import { createClient } from "@anam-ai/js-sdk";
2
3const res = await fetch("/api/anam-session", {
4 method: "POST",
5 headers: { "Content-Type": "application/json" },
6 body: JSON.stringify({ avatarId, agentId }),
7});
8const { sessionToken } = await res.json();
9
10const client = createClient(sessionToken);
11await client.streamToVideoElement("avatar-video");

Server Side

Client Side

1import { createClient } from
2 "@anam-ai/js-sdk";
3
4const res = await fetch(
5 "/api/anam-session", {
6 method: "POST",
7 headers: { "Content-Type":
8 "application/json" },
9 body: JSON.stringify({ avatarId,
10 agentId }),
11});
12const { sessionToken } = await res.json(
13 );
14
15const client = createClient(
16 sessionToken);
17await client.streamToVideoElement(
18 "avatar-video");

Architecture of Integration

Get started

Best Practices for ElevenAgents

Get started

Server-Side Implementatio

Get started

Client Side Implementation

Get started

How it works

Server-side architecture connects Anam's engine directly to ElevenLabs. Your client only uses the Anam SDK.

1

Server Authentication

Next.js API route fetches a signed ElevenLabs WebSocket URL and requests an Anam session token.

2

Engine WebSocket Connection

3

Client WebRTC Stream

4

ElevenLabs Processing

5

Avatar Generation

6

Unified Delivery

Real-time avatar generation workflow

How it works

Server-side architecture connects Anam's engine directly to ElevenLabs. Your client only uses the Anam SDK.

1

Server Authentication

2

Engine WebSocket Connection

3

Client WebRTC Stream

4

ElevenLabs Processing

5

Avatar Generation

6

Unified Delivery

Real-time avatar generation workflow
0123456789
0123456789

%

better overall experience than the nearest competitor

Overall experience

Naturalness

Responsiveness

Interruptibility

ANAM

Performance bar

TAVUS

Performance bar

HEYGEN

Performance bar

D-ID

Performance bar

CARA-3, our latest real-time avatar model, ranks #1 across all relevant metrics per Avatar Benchmark (2025) carried by Mabyduck.

0123456789
0123456789

%

better overall experience than the nearest competitor

Overall experience

Naturalness

Responsiveness

Interruptibility

ANAM

Performance bar

TAVUS

Performance bar

HEYGEN

Performance bar

D-ID

Performance bar

CARA-3, our latest real-time avatar model, ranks #1 across all relevant metrics per Avatar Benchmark (2025) carried by Mabyduck.

Powered by CARA. Our industry-leading real-time avatar model.

Industry lowest latency.
Our avatars respond faster than you blink.

33%

33%

33%

faster than next-best competitor

180MS

180MS

180MS

avg. agent response time

Real-time responses

Speed designed for real-time interactions means conversations feel seamless.

Support 70+ languages

Reach international markets over 70 languages covering 95% of the world, all with native voices.

Ultra-low latency

Real-time, everywhere. Anam models deliver 180ms latency.

Any character. Any style.
The perfect agent for your brand.

Enterprise-grade security and infrastructure for deploying Conversational Video Agents at scale.

Enterprise-level data protection

Photorealistic faces brought to life through natural motion and micro-expressions. No artificial gestures powered by our foundation video models

Zero Data Retention

Enterprise support and 99.9% uptime

Enterprise security and compliance icons

Enterprise-grade security and infrastructure for deploying Conversational Video Agents at scale.

Enterprise-level data protection

Photorealistic faces brought to life through natural motion and micro-expressions. No artificial gestures powered by our foundation video models

Zero Data Retention

Enterprise support and 99.9% uptime

Enterprise security and compliance icons

Integrate your agent today.
Use our widget or integrate using our API.

No-Code platform

API Integration

Desktop UI
API code snippet

No-Code platform

API Integration

Desktop UI
API code snippet

See how interactive avatars are disrupting industries

See how interactive avatars are disrupting industries

Working on a unique use case? Our team can help you out.

Built for enterprise scale.
High concurrency, ultra-low latency, and global distribution.

Concurrency that scales

Run thousands of simultaneous avatar sessions without degradation.

 Scalable concurrency diagram

Concurrency that scales

Run thousands of simultaneous avatar sessions without degradation.

 Scalable concurrency diagram

Concurrency that scales

Run thousands of simultaneous avatar sessions without degradation.

 Scalable concurrency diagram

Multiregion servers

Deploy sessions from the region closest to your users.

Multiregion server network

Multiregion servers

Deploy sessions from the region closest to your users.

Multiregion server network

Uptime Guaranteed

Your avatar is available when your users are.

99.9 percent uptime

Uptime Guaranteed

Your avatar is available when your users are.

99.9 percent uptime

Model Versioning

Pin to a specific avatar model version or promote to the latest flagship.

 Model version list

Model Versioning

Pin to a specific avatar model version or promote to the latest flagship.

 Model version list

Intelligent LLM Routing

Connect any LLM — GPT-4o, Claude, Mistral, or your own.

 LLM routing diagram

Intelligent LLM Routing

Connect any LLM — GPT-4o, Claude, Mistral, or your own.

 LLM routing diagram

Intelligent LLM Routing

Connect any LLM — GPT-4o, Claude, Mistral, or your own.

 LLM routing diagram

HIPPA and SOC II

Built for regulated industries. Anam is HIPAA compliant and SOC II certified.

HIPAA and SOC II compliance

HIPPA and SOC II

Built for regulated industries. Anam is HIPAA compliant and SOC II certified.

HIPAA and SOC II compliance

HIPPA and SOC II

Built for regulated industries. Anam is HIPAA compliant and SOC II certified.

HIPAA and SOC II compliance

We’re a research-led product company

Our research teams investigate the safety, inner workings, and societal impacts of AI models – so that artificial intelligence has a positive impact as it becomes increasingly capable.

Background pattern

Frequently asked questions

How does billing work?

You get billed by the second. Each plan includes a set amount of free minutes per month. Once you exceed your included minutes, you are charged per second of usage at your plan's per-minute rate and nothing more.

What counts as a minute?

Minutes are counted from the moment you start a conversation with an AI avatar (pressing "Start Chat" or initiating a session via the API). The timer runs for the entire duration of the session until you end it, regardless of whether you are actively speaking or not.

What happens to unused minutes?

Unused minutes do not roll over. If you have 250 free minutes on Explorer and only use 50, the remaining 200 minutes expire at the end of your billing period. Similarly, if you cancel or downgrade your plan, any unused minutes from your previous plan are not carried over to your new plan.

How do I cap my usage?

You can control your usage in two ways: set a spend cap in your Account Settings to limit how much you can be charged for extra minutes beyond your included allowance, or set conversation time limits programmatically using the SDK when initiating sessions.

What is a conversation limit?

The conversation limit is the maximum length of a single session with an AI avatar. Free and lower tier plans have time limits per conversation, while Growth and above have unlimited conversation length.

What is a simultaneous session?

A simultaneous session (also called concurrency) is the number of avatar conversations that can happen at the same time. As a general rule of thumb, we typically see a 200:1 ratio between daily active users and concurrent sessions needed. So if you expect 100 active users, you would typically need 1 simultaneous session slot.

What are custom avatars?

Custom avatars (One-Shot Avatars) are AI-powered video avatars you can create and use in your projects. Each plan includes a certain number of avatar slots that determine how many custom avatars you can have active at once.

How can I create a custom avatar?

You can upload an image via the lab or generate an image via text prompt to create a custom avatar. It takes <2 minutes to generate and come to life.

What is the "Remove Watermark" feature?

When embedding Anam avatars on your website, a watermark is displayed by default on the Widget. The "Remove watermark" option (available on Explorer and above) lets you hide this watermark for a cleaner branded experience.

Can I pause my subscription?

It's not possible to pause your paid subscription, but when you cancel your paid subscription, your account will be transferred to the free plan at the end of your current billing cycle. Unused minutes will not be carried over to the free plan.

What is Custom Voice Clone?

Custom voice clones allow you to create a unique AI voice that matches a specific person or style. You can upload voice samples to train a personalized voice for your avatar. This feature is available on Professional and Enterprise plans.