AIAP Tool Guides·Genspark·16:26

Genspark Image and Video Generation: Image Studio, Mixture of Agents, Automatic Prompt Expansion, and Veo 3 with Sound

Genspark guide presenter (likely Karan Rana) Marketing manager / YouTuber profile ("I am a youtuber"); solo screen-share on a Genspark Plus account

▶LMS lesson

The short version

All Agents > Image Studio: background removal, erase a person, magic redraw, restore old photos, image expand; Image Generation lets you pick a model or "mixture of agents", a style and an aspect ratio (0:01-0:06).
"A big Bengal tiger running across French streets", 3D, 16:9: four images from different models (mostly Flux, one Recraft) with each expanded prompt visible (0:06-0:10).
Generate Video: 9:16, 3-5 seconds, Veo 3 "with sound"; a Ferrari through Bombay streets in about two minutes (0:10-0:15).

At a glance, three clicks deep

Skim here first: the closed row is the glance, open is the study card with the key points and timestamps, and the ↓ link drops to that concept's full write-up below.

01Image and video generation with model choice and prompt expansionChoose a model or let Genspark pick ("mixture of agents"), set style and aspect ratio, write a short prompt;›

Choose a model or let Genspark pick ("mixture of agents"), set style and aspect ratio, write a short prompt; Genspark expands it and renders.

Image Studio (0:01-0:03)

Image generation (0:03-0:10)

Video (0:10-0:15)

↓ Full write-up of this concept

The concepts in full

Image and video generation with model choice and prompt expansion

how-to

No Midjourney syntax needed - the expanded prompt is shown and reusable.

The tiger prompt became a detailed 3D, photorealistic brief; outputs were 1360x768 (close to 16:9). He preferred Flux 1.1 and found Recraft dull and cartoonish; the streets looked more Italian than French. For video, the model menu listed mixture of agents, Veo 3 (adds sound effects and music), Kling, Luma, Hunyuan and Runway; the five-second Ferrari clip took about two minutes.

Do it in this order

01All Agents > Image Studio > image generation.
02Pick model or mixture of agents; choose style and aspect ratio.
03Prompt; compare outputs; read the expanded prompt; download.
04All Agents > Generate Video: set ratio, duration, model (e.g. Veo 3); prompt; wait; download.

Why it matters

Compare several models from one prompt.

This idea elsewhereparallelAI for Marketing: Genspark image-to-video product shot

Tools referenced

Tool	Coverage	Moment	Context
Genspark	demonstrated
Flux	demonstrated
Recraft	demonstrated
Veo 3	demonstrated
Imagen 4	mentioned
Kling	mentioned
Runway	mentioned
Midjourney	mentioned

Action items

Resources mentioned

Resources

docLesson page on the platform

Extraction notes

This page was built from an auto-generated transcript, which garbles product and people's names. Those were corrected silently in everything above and logged here for transparency. The warnings flag claims that were true on the recording day but change fast.

Transcript corrections applied

The transcript says	The trainer actually means
imagine 4	Imagen 4
clang	Kling
Hanyan	Hunyuan
gemini vo3	Veo 3

True on recording day — verify before relying

Vendor UI walkthrough: menu names, layout, plan limits and pricing change monthly - check the live product before following click paths.
Imagen 4 "just dropped a month back" (June 2025); the model lists and ~2-minute Veo 3 render are point-in-time (0:04-0:14).