const video = new AIVideo(article);

const video = new AIVideo(article);

const video = new AIVideo(article);

const video = new AIVideo(article);

await video.render({ language: "multilingual" });

await video.render({ language: "multilingual" });

await video.render({ language: "multilingual" });

await video.render({ language: "multilingual" });

AI-POWERED VIDEO CREATION

From words to WOW. Launch videos in minutes.

Create engaging, branded video content with AI-powered avatars, multilingual capabilities, and complete editorial control. Perfect for newsrooms and content publishers.

AI Video Preview

30+

Languages

85%

Time Saved

AI-Powered

Transform Text to Professional Video

See how easy it is to create studio-quality videos from your content in just minutes.

ContentAvatarStyleSettingsGenerate

~ min video

Enter Your Content

Paste your article, script or prompt for the AI to turn into a video.

Tax Avoidance Among Tech Billionaires: The Nvidia Case

Today, we're uncovering the story of a tech titan and how billions in taxes are being avoided. Meet Jensen Huang, the mastermind behind Nvidia and one of the wealthiest individuals in the United States.

Jensen Huang, the chief executive of Nvidia, is worth an astonishing $127 billion. As the leader of the company that powers much of artificial intelligence, Huang is celebrated for his engineering genius. But he's also making headlines for a different reason—his ability to shield billions of his fortune from taxes.

At 61, Huang has employed strategies that could save his family an estimated $8 billion in federal estate taxes. This approach places him among a growing group of ultrawealthy individuals who use financial maneuvers to bypass significant tax obligations.

Huang is not alone. Executives from companies like Meta, Google, and Blackstone have shifted billions into financial vehicles designed to avoid the federal estate tax. Despite being intended for only the wealthiest estates, the tax has been steadily weakened over the years.

Here's the bigger picture: Since 2000, the wealth of the richest Americans has roughly quadrupled, yet revenue from the estate tax has remained nearly unchanged. If the tax had kept pace, it could have brought in $120 billion last year. Instead, it raised just a fraction of that amount.

This story is a reminder of how the ultrawealthy are reshaping financial systems, often leaving taxpayers to wonder about the fairness of it all.

Thanks for joining us, and we'll see you next time with more insights!

~600 characters

Choose from Our AI Avatar Library

Every presenter below is enabled on the platform today, each in several looks, and each able to speak your script in any of the supported languages.

Gergana Borislavova

7 looks

Stefani Ivanova

1 look

Armando

2 looks

Bahar

5 looks

Diana

3 looks

Florin

4 looks

Hada

8 looks

Javi

9 looks

Judith

4 looks

Shirley

4 looks

Emery

5 looks

Bryce

5 looks

Need a custom avatar that matches your brand? We also offer avatar cloning services.

Capability demo

One Video, Multiple Languages

Reach global audiences with perfectly lip-synced videos in any language you need, all from a single source.

🇺🇸English Version

This video has been automatically translated and lip-synced to Englishwhile preserving the presenter's gestures, intonation, and expressions.

Our technology supports over 70 languages with more being added regularly. Custom language packs are available for enterprise clients.

Real customer output

Published articles, turned into video

Every video below was generated by a publisher from one of their own articles. Nothing here is a demo reel. The tab is the language of the video; the headline is the article it was made from, which is often in another language.

България привлича чуждестранни инвеститори с достъпния си имотен пазар

ForbesTikTok Video1080×1920

Близо 60% от българските ученици са функционално неграмотни по четене и математика

MaikoMilaInstagram Reel1080×1920

Джо Роугън е притеснен за завръщането на Илия Топурия след загубата от Гейджи

MMA BGInstagram Reel1080×1920

Европейските фермери искат спешни мерки след горещите вълни

AgriBGGeneric Video1920×1080

България остава ключов пазар за гръцките костилкови плодове

AgriBGInstagram Reel1080×1920

How to Buy a Hardwood Slab in Europe

KORENAYouTube Video1920×1080

If you would rather build it

Compared with Schibsted's open-source Videofy

Schibsted has open-sourced a working version of this idea, and it is a good piece of engineering. If your team wants to run the pipeline itself, start there rather than with us. Here is an honest account of where the two differ — checked against the repository, not against its README.

	Videofy Minimal	MediaThrive
Pipeline	OpenAI writes the manuscript and matches media, ElevenLabs narrates, Remotion renders	The same class of pipeline, run for you as a service
Where it runs	On a laptop, or in local containers, with your own API keys	Hosted and monitored, with queueing and retries
Built for	Digital signage screens	Newsroom web and social
Output sizes	Vertical and horizontal	Seven presets across four sizes, from 9:16 to 16:9
Branding	Per-brand JSON files: logo, intro, wipe, outro, colours, voice, prompts	The same controls, managed per team inside the product
On-screen text	Per-segment text lines, styled and placed per brand	Word-timed captions burned into the frame, in six preset styles
Presenter	Plays reporter intro and outro clips you have recorded yourself	A generated AI presenter, lip-synced to the narration
Getting articles in	Reuters and AP fetchers, plus a generic HTML one for testing; "each brand/newsroom should build a proper fetcher integration"	A monitoring pipeline that already watches your sources

Checked against github.com/schibsted/videofy_minimal on 12 September 2026. That repository is the public subset of Videofy; its README says it leaves out most of the integrations and infrastructure Schibsted runs internally, so nothing here is a claim about Schibsted's own product.

Before you sign anything

What this does not do

You would find all four of these out in your first week. Better to read them now.

Rendering takes minutes, not seconds

Across our last 45 production renders the median was under four minutes, and nine in ten finished inside ten. That is roughly two and a half times the length of the finished video. Long pieces cost more: the three biggest videos we have rendered, at thirty to forty scenes, took 29, 44 and 84 minutes.

The input is a published article, not a prompt

Every video we have ever produced was made from an article in the system — there is no prompt box. If the story has not been written and published yet, there is nothing to turn into video.

Captions are burned in

Captions are rendered into the frame so the video reads correctly in a muted feed. We do not hand back a separate subtitle file, so you cannot switch them off or re-style them after the render without rendering again.

The pictures come from your story

Visuals are drawn from the article and your own library first. An article that ships with one thumbnail gives the video less to work with than one with a full photo set, and it shows.

Also in MediaThrive

One platform. All your content needs.

AI AudioArticle narration AI PodcastsAutomated shows AI TextArticle generation Media MonitoringNews intelligence

Evaluating Synthesia, Fliki, or Pictory?

See why publishers choose MediaThrive — a complete newsroom OS — over single-purpose video tools.

MediaThrive vs Synthesia MediaThrive vs Fliki MediaThrive vs Pictory

Frequently Asked Questions

Find answers to common questions about our AI video technology, licensing, and implementation.

Legal & Compliance

Usage & Technical

Can't find what you're looking for?

Contact our team for more information →

Transform Your Content Strategy with AI Video

Join leading media companies who are creating professional news videos at scale. Start producing engaging video content in minutes, not days.

Start Free Trial

Schedule a Demo

No credit card required. 14-day free trial with 50,000 credits.