Imagen

AI Avatar API — Generate Avatars from a Selfie

Generate styled AI avatars from a single selfie with one POST request. Face identity stays, the style follows your prompt. Image plans from $21/month for 3,250 API calls.

Get API Key — $21/Month API Documentation

Last updated Sep 27, 2026 · By ModelsLab Engineering

Which AI avatar and talking-head services offer API access?

"AI avatar API" means two different products. A talking-head API turns a photo or a trained digital twin into a video of a person speaking your script: HeyGen, D-ID, Synthesia, Tavus and Akool all sell one, from about $0.60 to $2.31 a minute of video. HeyGen's API is pay-as-you-go with no plan required; D-ID, Synthesia and Tavus bundle API minutes into monthly plans.

An avatar image API turns a selfie into a styled portrait, which is what this page covers: ModelsLab face_gen returns one avatar per call from plans at $21/month (3,250 calls). ModelsLab also has talking-head models on the same key, OmniHuman 1.5 at $0.14 a second ($8.40 a minute) and lipsync-2 at $0.07 a second for syncing new audio onto an existing video. Per minute, those cost more than the dedicated avatar vendors, so for long presenter videos HeyGen or D-ID is cheaper.

Talking-Head Avatar API Prices, September 2026

Price per minute of generated video from each vendor's own pricing page. Where a vendor sells plans, the figure is the plan price divided by its included minutes.

Provider	Product	About per minute of video	Plan needed for the API
HeyGen	Avatar III digital twin / photo avatar	$0.60 / $0.99	None, pay as you go
HeyGen	Avatar IV photo avatar	$2.31	None, pay as you go
D-ID	Photo avatar video	About $0.99 to $1.13	Build $18/month (16 min) to Scale $297/month (300 min)
Synthesia	Avatar video	About $2.23	Creator, $89/month or $804/year
Tavus	Replica video	$0.90 to $1 over the plan	Starter $59/month (10 min)
Akool	Talking avatar, 1080p	About $1.02	Pro Max (first plan with API access)
ModelsLab	OmniHuman 1.5 (omni-human-1.5)	$8.40	Any plan, from $21/month
ModelsLab	lipsync-2, new audio on a video	$4.20	Any plan, from $21/month

Prices checked September 27, 2026 on help.heygen.com (HeyGen API pricing explained), d-id.com/pricing/api, synthesia.io/pricing, tavus.io/pricing and akool.com/api-pricing ($0.034 per credit on Pro Max). ModelsLab prices are from its model catalogue: omni-human-1.5 $0.14 per second, omni-human $0.168 per second, lipsync-2 $0.07 per second.

What the AI Avatar API Does

One Selfie In, a Styled Avatar Out

The AI avatar API takes a single face photo and a text prompt and returns an image that keeps the identity of the face while rendering it in the style you describe — a studio headshot, an anime character, a 3D render, a comic panel. There is no template library to browse and no multi-photo training step; the face is read from the reference image at request time.

It runs on the same REST API and the same key as the rest of the ModelsLab image catalogue, so an avatar can be upscaled, relit, animated or lip-synced by passing its URL to the next endpoint.

What a Request Looks Like

POST https://modelslab.com/api/v6/image_editing/face_gen with `face_image` set to the selfie URL and `prompt` describing the style. Optional fields cover `negative_prompt`, `width`, `height`, `samples` (up to 2 per call), `seed` for reproducibility, a `style` preset (`realistic`, `anime`, `3d`, `comic`, `chibi`) and a `webhook` URL for async delivery.

Face identity preserved from one reference photo
Style controlled by prompt or the style preset
Up to 2 samples per request, optional width and height
Hosted image URLs in the response — no base64 handling
Seed parameter for repeatable results
Same key covers upscaling, relighting and video

AI Avatar API quick start

One POST to /api/v6/image_editing/face_gen with a selfie URL and a style prompt. Returns a hosted URL for the rendered avatar.

Python — generate an avatar from a selfie

Python

1import requests
2
3response = requests.post(
4    "https://modelslab.com/api/v6/image_editing/face_gen",
5    json={
6        "key": "YOUR_MODELSLAB_API_KEY",
7        "face_image": "https://example.com/selfie.jpg",
8        "prompt": "professional headshot, studio lighting, sharp focus",
9        "negative_prompt": "blurry, distorted, low quality",
10        "style": "realistic",
11        "width": 1024,
12        "height": 1024,
13        "samples": 1,
14    },
15)
16
17print(response.json()["output"][0])

JavaScript — anime avatar with a fixed seed

JavaScript

1const response = await fetch(
2  'https://modelslab.com/api/v6/image_editing/face_gen',
3  {
4    method: 'POST',
5    headers: { 'Content-Type': 'application/json' },
6    body: JSON.stringify({
7      key: 'YOUR_MODELSLAB_API_KEY',
8      face_image: 'https://example.com/selfie.jpg',
9      prompt: 'anime avatar, cel shading, soft studio light',
10      style: 'anime',
11      seed: 42,
12      samples: 1,
13    }),
14  },
15);
16
17const data = await response.json();
18console.log(data.output[0]);

cURL — quick test from the terminal

bash

1curl -X POST 'https://modelslab.com/api/v6/image_editing/face_gen' \
2  -H 'Content-Type: application/json' \
3  -d '{
4    "key": "YOUR_MODELSLAB_API_KEY",
5    "face_image": "https://example.com/selfie.jpg",
6    "prompt": "corporate portrait, neutral background",
7    "style": "realistic",
8    "samples": 1
9  }'

Avatar Endpoints on the Same Key

face_gen is the avatar endpoint. The neighbours below cover the steps before and after it.

Endpoint	What it does	Typical use
POST /api/v6/image_editing/face_gen	Styled avatar from one selfie	Profile pictures, characters, headshots
POST /api/v6/image_editing/head_shot	Same endpoint, headshot-oriented alias	Corporate and team headshots
POST /api/v6/image_editing/super_resolution	Upscale the avatar	Print and large-format use
POST /api/v7/video-fusion/image-to-video	Animate the avatar	Intro clips, social loops
POST /api/v7/video-fusion/lip-sync	Make the avatar speak an audio track	Talking-head explainers

Every endpoint accepts and returns hosted image URLs, so the output of one is the input of the next.

Ship Avatars in 3 Steps

From API key to a styled avatar in one request.

STEP 01

Step 1: Get Your API Key

Create a ModelsLab account, subscribe to an image plan from $21/month for 3,250 API calls, and copy your key from the dashboard.

STEP 02

Step 2: Send the Selfie and a Style Prompt

POST to /api/v6/image_editing/face_gen with face_image set to a hosted selfie URL and prompt describing the look. A straight-on, well-lit photo gives the most faithful identity.

STEP 03

Step 3: Use, Upscale or Animate the Result

The response carries a hosted image URL. Serve it directly, upscale it with super_resolution, or pass it to image-to-video or lip-sync to make the avatar move and speak.

Get API Key

Related image API guides

Image Editing API

Inpainting, object removal, background replacement and upscaling on the same key.

Image Generation API

10,000+ models for text-to-image and image-to-image at a flat $0.0047.

Lip Sync API

Make a generated avatar speak any audio track with matched mouth movement.

Consistent Characters Across Many Images

face_gen is the fastest route to one avatar from one photo. When a product needs the same character across dozens of scenes — a game, a brand mascot, a virtual presenter — train a LoRA on a small photo set instead and generate from that model with the text-to-image endpoint. Both paths run on the same plan and key; the difference is setup time against consistency.

Pricing

Image plans start at $21/month for 3,250 API calls and $47/month for 10,000. Avatar generation on self-hosted models is one call per image and is covered by the plan. The $149/month Open Source Unlimited plan removes the per-call limit on the open-weight image models ModelsLab self-hosts. There is no separate avatar fee.

Why Developers Use the ModelsLab Avatar API

Key advantages that set us apart

One selfie in, styled avatar out — no training step

Face identity preserved, style set by prompt

Five style presets: realistic, anime, 3D, comic, chibi

Hosted image URLs back in seconds

Seed parameter for repeatable output

Chain to upscale, relight, animate or lip sync

Image plans from $21/month for 3,250 calls

One API key across image, video, speech and LLM

Our Popular Use Cases

What teams build with the AI avatar API:

Offer a generated avatar during onboarding from the photo the user already uploaded. One request per style option.

AI Avatar API FAQ

POST https://modelslab.com/api/v6/image_editing/face_gen with face_image set to a selfie URL and prompt describing the style. The response carries a hosted URL for the avatar, or an id to fetch when the request is queued.

Image plans start at $21/month for 3,250 API calls and $47/month for 10,000. Avatar generation on self-hosted models costs one call per image and is covered by your plan.

Any style you can describe in the prompt, plus five presets: realistic, anime, 3D, comic and chibi. Face identity is kept from the reference selfie while the style follows the prompt.

No. face_gen works from a single selfie. For a fixed character across many images, train a LoRA on a small photo set and generate avatars from that model instead.

ModelsLab is a paid service. Image plans start at $21/month for 3,250 API calls. Creating an account is free, but generating avatars requires an active plan.

Yes. Pass the avatar URL to /api/v7/video-fusion/image-to-video to animate it, or to /api/v7/video-fusion/lip-sync with an audio track to make it speak. All of it runs on the same API key.

Your Data is Secure: GDPR Compliant AI Services

ModelsLab GDPR Compliance Certification Badge

GDPR Compliant

Plans from $21 a month

Basic is $21/month and Standard $47/month; self-hosted image models cost $0.0047 per image from that allowance. Open Source Unlimited is $149/month with no per-image charge on self-hosted models. Every plan covers Flux, SDXL and 10,000+ community models. Cancel anytime.

100% refund policy on monthly & yearly plans — cancel anytime

Contact Sales

Best Value

Open Source Unlimited

Mission-Critical

$149/month

100% refund policy · cancel anytime

Unlimited Open Source Models

100% refund policy

24x7 Support

15 parallel generations

Access to all APIs

Unlimited generations on all open-source models

For mission critical workloads

Add Team Members

Priority GPU Clusters

Standard

Production

$47/month

100% refund policy · cancel anytime

Moderate Traffic

100% refund policy

Priority Developer Support

10 concurrent API requests

For Production workloads

API access to all models

Prototype

Basic

Prototype

$21/month

100% refund policy · cancel anytime

Moderate Traffic

100% refund policy

Developer Support via Discord/Email

5 concurrent API requests

API access to all models

Shared GPU

Get Expert Support in Seconds

We're Here to Help.

Want to know more? You can email us anytime at support@modelslab.com

View Docs

POST https://modelslab.com/api/v6/image_editing/face_gen with `face_image` set to a selfie URL and `prompt` describing the style. The response carries a hosted URL for the avatar, or an id to fetch when the request is queued.

Image plans start at $21/month for 3,250 API calls and $47/month for 10,000. Avatar generation on self-hosted models costs one call per image and is covered by your plan.

Any style you can describe in the prompt: realistic headshots, anime, 3D render, illustration, pixel art. Face identity is kept from the reference selfie while the style follows the prompt.

No. face_gen works from a single selfie. For a fixed character across many images, train a LoRA on a small photo set instead and generate avatars from that model.

ModelsLab is a paid service. Image plans start at $21/month for 3,250 API calls. Creating an account is free, but generating avatars requires an active plan.

Yes. Pass the avatar URL to /api/v7/video-fusion/image-to-video to animate it, or to /api/v7/video-fusion/lip-sync with an audio track to make it speak. All of it runs on the same API key.

Talking-head APIs, which turn a photo or digital twin into a speaking presenter video: HeyGen (pay as you go, $0.60 to $2.31 a minute by avatar type), D-ID (from $18/month for 16 minutes), Synthesia (from the $89/month Creator plan), Tavus (from $59/month) and Akool (about $1.02 a minute). ModelsLab has OmniHuman 1.5 at $0.14 a second and lipsync-2 at $0.07 a second on the same key as its avatar image API, face_gen, which turns a selfie into a styled portrait on plans from $21/month. Prices checked 2026-09-27.