InVideo launches v3.0 to generate full videos from text prompts

Most "AI video generators" on the market are still glorified slideshow makers, stringing together awkward stock footage with robotic voiceovers. InVideo wants to change that narrative.

The Tiger Global-backed editing platform just rolled out InVideo v3.0, introducing prompt-based generative AI creation that generates original footage from scratch.

Moving past stock footage pipelines

Instead of stitching together licensed stock clips like its earlier version, v3.0 uses text prompts to output videos across styles like live-action, animation, and anime. Users can export directly into formats tailored for YouTube, vertical Shorts or Reels, and LinkedIn.

If a scene looks off, you do not have to scrap the project. The editor lets users prompt for precise tweaks or additions to specific sections of a clip.

InVideo did not build its own foundation model. Instead, it relies on a multi-model pipeline that parses prompts and calls different models to generate the final output. While co-founder and CEO Sanket Shah said there is no technical limit to video length, generation relies strictly on paid credits, with no free tier available.

The steep cost of generative minutes

AI video generation remains brutally expensive, and InVideo is passing that cost directly to users. The company launched a Generative Plan priced at $120 per month, which yields just 15 minutes of generated video. Additional time costs $8 to $10 per minute.

The output is also far from seamless. Sample clips on the platform's site reveal jarring shifts in visual style and quality midway through a video, an issue the company says it plans to address over time.

Still, InVideo is negotiating from solid footing. The platform counts 4 million monthly active users, produced 7 million videos in the last 30 days, and is pacing toward $50 million in revenue this year. Shah noted that out of the $35 million raised in 2021, the startup still holds $25 million in cash.

Why it matters

For independent creators and small businesses, the promise of text-to-video is huge, but at $8 a minute, the margin for error is razor-thin. Casual creators simply cannot afford to iterate when a single botched prompt burns through real cash.

Strategically, InVideo is taking a smart positioning angle against rivals like Captions, D-ID, and Lightricks. While others chase enterprise marketing teams and agency budgets, InVideo is sticking with consumer-friendly interfaces.

However, stitching together third-party models comes with serious vulnerabilities. InVideo does not control model consistency, leading to visual flickering and sudden style jumps. It also remains hostage to third-party compute margins. Unless InVideo can tame these mid-video style collapses and bring per-minute costs down, everyday creators might stick with traditional editing tools a while longer.