One face oil photo is the whole shoot
GESTEL learns your brand, reads your product, and makes the photos and clips that look like you.
This is how it goes on the canvas
Say it on the left, watch the shots grow on the right. Call a product with @, pin a brand tone, dress a model.
@lipstick@Late AutumnMake a packshot, a scene shot and a model shot in our brand tone.Keeping the engraving on the case and the lip colour as they are: one packshot, a scene among persimmons and dry leaves, a model shot in hand, a mirror flat lay and a detail, and a thumbnail — six shots.
Made 6 shots · 6 images used
Know the brand, read the product, then shoot
Start with one product photo and one line. GESTEL reads the brand and the product first, then makes the shots you need in that tone.
The product never changes
The product is read first and the instruction followed exactly. However the scene changes, the shape, colour and label text stay as they are in the photo.

@lipstickOn old wood among ripe persimmons and dry leaves, late-afternoon sun through a window.
@cardiganOn a model in a silver-grass field at sunset, the sea and an island behind.
@face oilOn a mossy stone in the woods, sun through the leaves, a little dew.
@necklaceOver a black turtleneck, the face cropped at the lips, side light from a window.How it works
First it learns the brand, then it reads the product, and only then does it make the photos and clips.
Prepared before the shoot
One good shot matters less than the same product coming out in the same brand tone every time. That is why the brand and the product are read first.
Start faster from a scene
Scenes gathered by category, each shot the way that category sells. Pick one and remake it with your product.
AI-generated examples
Any product, true to its category
Skincare, bags, jewelry, watches, candles, desserts, appliances — each category has its own staging that sells, learned separately.
AI-generated examples
One subscription, every model
Every image and video model on one plan. GESTEL picks the model for the job, or you pick it yourself.

Google
- Nano Banana 2Semantic editing, no masks needed.
- Nano Banana ProTop-tier editing with character consistency.
- Veo 3.1videoState-of-the-art video generation.
- Veo 3.1 FastvideoThe faster version.
- Veo 3.1 LitevideoBalances everyday utility with pro capability.
- Gemini Omni FlashvideoReal-world motion with synced audio.

OpenAI
- GPT Image 2.5 FlareFast, natural-looking default image model.
- GPT Image 2.5 SunburstPrecision model with extra fidelity on fine detail.
- GPT Image 2Fine-grained edits at draft quality.

ByteDance
- Seedream 5 ProFlagship, strong on text and dense layouts.
- Seedream 5 LiteFast editing from multiple inputs.
- Seedance 2.5videoStays coherent across a long shot.
- Seedance 2.0videoTop video model, with camera control and audio.
- Seedance 2.0 MinivideoA faster, cheaper version.

Kling
- Kling 3videoCinematic visuals with native audio.
- Kling 3 4KvideoNative 4K, no upscaling.
- Kling O3 ProvideoBridges a start frame to an end frame.
- Kling 2.6 ProvideoPremium cinematic video with native audio.

Black Forest Labs
- FLUX 2 MaxThe most realistic and precise tier.
- FLUX 2 ProProduction-grade editing, no configuration.
- FLUX 2 FlexAdjustable steps and typography fixes.
- FLUX.2 [klein] 9BPrecise edits from language and hex colors.
- FLUX 3videoFrontier video with natural motion.
- FLUX 3 DraftvideoFast, low-cost draft previews.

Alibaba
- Qwen Image 3Strong at text rendering and prompt adherence.
- Qwen Image MaxMore realistic, natural images.
- Wan 3.0 PrimevideoKeeps the source's look as it moves.

xAI
- Grok Imagine 2Generates and edits images.
- Grok Video 1.5videoVideo with audio.

MiniMax
- MiniMax H3videoTransitions from a first and last frame.
- MiniMax H3 MaxvideoTuned for prompt adherence and aesthetics.

Luma
- Luma Uni 1Restyles while keeping the composition.
- Luma Ray 3.2videoCinematic video, with seamless loops.

Lightricks
- LTX 2.5 FastvideoOpen-source, audio in one pass, built for iteration.

Others
- MAI Image 2.5Photoreal images with pixel-level control.
- Ideogram V4Reworks to your prompt, keeping the structure.
- Reve 2.1Accurate layout and text rendering.
- HiDream O1Generation, editing and custom subjects in one model.
- Hunyuan Image 3Instruction editing with internal reasoning.
- Z-Image TurboA very fast lightweight model.
- PixVerse V6videoThe latest V6 model.
- Vidu Q3videoThe latest Q3 Pro model.
- Marey 1.5videoTrained only on licensed data.
One product photo. Upload it now.
A studio shoot costs thousands. Make your product shots in GESTEL instead.













































































