GA Decart

Decart Realtime (Lucy)

Live video-to-video editing and restyling over WebRTC: send a webcam stream plus a text prompt or reference image and get the transformed stream back. Also offers a realtime world model preview (Oasis 3).

Est. per minute$0.60 - 2.40
1 high-severity warning

Overview

Best for: Live AR-style filters, virtual try-on, and creative live streams.

At a glance

Session $/min$0.60
Entry plan $/mo$0
Free tierYes
Resolution p720
WebRTCYes
Self-hostNo
Open weightsNo

Price is Lucy Restyle 2 ($0.01/s); Lucy 2.5 is $1.20/min, fast mode $2.40/min. Billed at 720p; 1080p option exists. Free credit amount not stated.

Audio in

None (video only)

Audio out

None

Languages

Prompts in text; language support not documented.

Latency

No number on the model page; fast mode described as lower latency and higher throughput.

Regions

Not captured; network requirements page exists.

Compliance

Not verified.

Features

  • realtime video-to-video
  • text prompt and reference image control
  • short-lived client tokens
  • SDKs for JS, Python, Swift, Android
  • auto-reconnect (5 retries)
  • 1080p option on connect

Pricing

WhatPriceUnitNotes
Lucy 2.5 realtime$0.02 ($0.04 fast mode)per second of active generation720p
Lucy VTON 3.6 / 3.5 realtime$0.02 ($0.04 fast for 3.5)per second720p
Lucy Restyle 2 realtime$0.01per second720p
Oasis 3 Preview realtime$0.02per second720p; enterprise pricing available
How the per-minute estimate was worked out

$0.01/s Restyle = $0.60/min up to $0.04/s fast mode = $2.40/min; Lucy 2.5 standard $1.20/min

Free tier: New accounts get free credits (amount not stated). No subscription or minimum spend.

Source: docs.platform.decart.ai

Setup

  1. Create an API key on platform.decart.ai.
  2. Add a server route that mints short-lived client tokens restricted to the models you use.
  3. npm install @decartai/sdk in the browser app.
  4. Capture the webcam at the model's fps/size and call client.realtime.connect with a prompt.
  5. Close the connection when the tab is hidden to stop billing.

Endpoint

WebRTC via @decartai/sdk (client.realtime.connect)

Authentication

API key server-side; short-lived client tokens in the browser

Quick start javascript

import { createDecartClient, models } from "@decartai/sdk";

// Browser: token comes from your server (serverClient.tokens.create)
const client = createDecartClient({
  apiKeyProvider: async () => {
    const r = await fetch("/api/realtime-token", { method: "POST" });
    return (await r.json()).apiKey;
  },
});

const model = models.realtime("lucy-2.5");
const cam = await navigator.mediaDevices.getUserMedia({
  video: { frameRate: model.fps, width: model.width, height: model.height },
});

const rt = await client.realtime.connect(cam, {
  model,
  onRemoteStream: (s) => { document.getElementById("out").srcObject = s; },
  initialState: { prompt: { text: "Turn me into a claymation character" } },
});

// Server (Express):
// const t = await serverClient.tokens.create({ expiresIn: 300,
//   allowedModels: ["lucy-2.5"] }); res.json(t);

Written from the current docs. Check the vendor's SDK version before you ship.

Warnings

Expensive per hour

$0.02/s is $72 per active hour per stream at standard speed and $144 in fast mode. Put a hard session cap and stop on visibility change.

Never ship your API key to the browser

Use server-minted client tokens with allowedModels and short expiry; the SDK refetches them on reconnect.

720p is the billed baseline

Pricing is quoted at 720p; marketing mentions 1080p. Check how 1080p is billed before enabling it.

Deepfake rules apply

Character swap and face edits on live video fall under deepfake disclosure rules (e.g. EU AI Act Article 50); label output and get consent for real people.

Model names change fast

Lucy 2.1 is already 'previous generation'; read fps/size from the model config instead of hardcoding.

Plus 9 warnings that apply to all avatars and live video APIs. See category warnings.

Limits

  • Concurrent realtime sessions limited by account quota (Get Realtime Quota endpoint)
  • Client tokens expire 60 s after minting by default

Models and products

NameStatusNotes
lucy-2.5GADefault realtime editing model: add/replace/remove objects, swap characters, restyle; 1280x720 landscape or portrait; fast mode at 2x price.
lucy-restyle-2GAStyle transfer at half the price.
lucy-vton-3.6 / 3.5GARealtime virtual try-on.
oasis-3 (preview)PreviewRealtime world model; enterprise pricing available.
lucy-2.1GAPrevious generation.

Docs and sources

Docs

Sources used

Not fully verified

Latency numbers, free credit amount, concurrency quota defaults, 1080p billing.

Similar avatars + video APIs

Spotted a wrong price or a dead link?