Lovart “Talking” Video Feature​

​

🔥 What’s New – Background of This Collaboration​

​

Lovart has just launched a groundbreaking AI video feature — Talking.​

It’s not just another talking-head tool — it’s the First AI system that can generate long-form, dual-person podcast-style videos with unmatched visual consistency and precise lip-syncing.​

While most AI tools are still stuck in short solo clips, Lovart takes it further: you get stable visuals, auto-generated scripts, and natural back-and-forth conversations — all fully AI-generated.​

This isn’t just a new feature — it’s a new content format.​

Perfect for testing, reacting, memeing, or even breaking down serious trends and hot takes.​

​

Sample Outputs – What It Looks Like​

​

​🔥 What’s New – Background of This Collaboration

Lovart has just launched a groundbreaking AI video feature — Talking. It’s not just another talking-head tool — it’s the First AI system that can generate long-form, dual-person podcast-style videos with unmatched visual consistency and precise lip-syncing.

While most AI tools are still stuck in short solo clips, Lovart takes it further: you get stable visuals, auto-generated scripts, and natural back-and-forth conversations — all fully AI-generated.

This isn’t just a new feature — it’s a new content format.

Perfect for testing, reacting, memeing, or even breaking down serious trends and hot takes.

Sample Outputs – What It Looks Like

Preview what Talking can do: high-quality samples showing dual-person dialogues, stable visuals, and synced lip movement – all generated with Lovart.

【Generate a podcast where two slices of sourdough bread realize they’re being eaten alive and start discussing free will.】

Talking – what it does

• Generate Single- or dual-person videos, up to 5 minutes, in one continuous shot.
• Fully customizable: set your Topic, script, characters, style, and scene — great for podcasts, interviews, explainers, product launches, or character dialogue.
• If you don’t provide content, Lovart handles it for you: → generates script → turns it into audio → picks a starting image → syncs lip movement → adds subtitles

How It’s Triggered

Just describe something Talking-related in your prompt, and Lovart will switch to Talking mode. Keywords like: Talk, introduce, explain, describe, discuss, express, present, debate, etc.

Popular use cases & examples

1. Give a topic, let Lovart improvise → ‘Generate a podcast video of a host talking about Google I/O 2025.’

2. Describe characters, setting, and topic in detail → ‘Create a 90-second French Revolution-era video in oil painting style. King Louis XVI is holding his own severed head in his arms. The head is talking calmly in a podcast-style conversation with Marie Antoinette, who is sipping tea and unbothered. The topic: ‘Was the revolution actually a design problem?’

3. Provide your own full script → Lovart will directly use your words to generate voice + video (see detailed example above).
4. Upload an image to define how the host/guest should look or what the background should be → ‘This is Melvin, CEO of Lovart. Make a podcast where he gets interviewed.’
5. Use a custom image as the first frame → ‘Start with this image in pop art style and make a video explaining pop art’s history.’

✅ All of the above can be mixed & matched. It’s flexible. It’s fast. It’s fun.

How to write a good prompt

Recommended Prompt Structure:

1. Start with:Generate a video
2. Specify character(s): e.g. One host, A girl, Donald Trump, Two hosts, A boy and a girl, Donald Trump and Elon Musk
3. Add a talking intent: e.g. Talk, introduce, discuss, explain, debate, describe, state
4. Add a topic: → Keep it simple (E.g. “About AI and design”) or go detailed (“explaining the impact of generative AI on creative workflows”).

Optional Controls (Highly Recommended):

• Video length: Default is ~90–150s (resource intensive). You can say: ‘keep it under 60 seconds’ or ‘around 90s’
• Subtitles: Added by default. To turn off, include: ‘do not add subtitles to the final video’
• Voice gender: You can specify: ‘Left character has a male voice, right character has a female voice’
• Aspect ratio: Default is 16:9. You can request: ‘make it vertical (9:16)’ or ‘square (1:1)’

Usage Notes

Pet-style characters (e.g. Cats, dogs) are currently Not recommended for Talking — lip-sync quality may be low.

Best results come from Characters with clear facial features: human-like characters, stylized 2D/3D personas, humanoid puppets/animals.

​