Sogni Photobooth for Web

Sogni Photobooth is the easiest way to turn a selfie into stylized AI art. Snap or upload a photo, pick a vibe, and the Sogni Supernet generates portraits, character variations, and short videos in seconds — no model downloads, no GPU, no install. It runs on the same Supernet that powers every Sogni app, so Photobooth is also the reference build for what's possible end-to-end.

#Launch Photobooth →

#What it does

• Identity-preserving portraits. Keep your face; transform everything else — outfit, era, medium, world.
• 200+ curated style presets across 16 vibe categories, with seasonal rotations.
• Personalize. Type a theme like "world-famous paintings" or "Singaporean pop-culture parody" and Photobooth generates a custom series of prompts on the fly.
• Group photos. Multi-face detection that auto-routes you to identity-preserving models when there's more than one person in frame.
• AI editing & transformations. Turn yourself into Lego, claymation, Pixar, pixel art, Ukiyo-e, neon — or add cats, hats, glasses, and Viking helmets.
• Camera-angle remapping that re-renders the same subject from any of 96 angle combinations, plus pre-built multi-angle character sheets.
• Seven video styles — Emoji Video, Motion Transfer, 360 Camera, Replace Subject, Sound to Video, Batch Transitions, Prompt Video — powered by Wan 2.2 and LTX-2.3.
• Dance Montage recipe. Drive your portrait with a reference dance video and ACE-Step-generated music.
• Batch up to 256 concurrent generations on desktop (16 on mobile) — a single tap fills a whole gallery.
• Frontier likeness via OpenAI GPT Image 2, wired in alongside Sogni's own SDXL/Flux/Qwen models.
• Event & kiosk mode with QR-watermarked outputs, shared galleries, and physical print fulfilment.

Photobooth is built on the Sogni Client SDK — the same SDK that powers every Sogni app on the Supernet — and is published as an open-source reference integration.

#The style library

Photobooth ships a hand-tuned library of ~200 style presets organized into 16 active categories. Each preset has a preview thumbnail in the picker so you can see the look before you generate.

Seasonal collections (Halloween, Christmas / Winter) rotate in around their events and disappear again when the season ends, so the picker stays focused.

#Personalize: generate your own vibe series

Don't see the exact look you want? Open Personalize and describe a theme in plain English:

• "Parody of world-famous paintings and art"
• "Photobooth themes for a 12-year-old boy's birthday party"
• "Famous magazine covers with titles"
• "Singaporean cultural landmarks and pop-culture parody"
• "Album covers from the 70s, 80s, 90s, 2000s"

Photobooth expands your input into up to 16 fully-formed prompts — each with its own name and styling — and drops them into a new Personalized category in your picker, ready to generate against any model.

Behind the scenes, Personalize defaults to GPT Image 2 for frontier likeness preservation, but you can swap it to Stable Diffusion or Qwen Image Edit if you want a different aesthetic. Personalized vibes persist per wallet (logged-in) or per session (demo), and at branded events they're shared across every attendee so a whole party can riff on the same custom theme.

#Group photos

Photobooth detects multiple faces in your input and, if your selected model is tuned for single portraits, offers to switch you to a context-aware model that preserves every identity in the frame — typically Qwen Image Edit 2511 or Flux.2 Dev. From there, every style preset, edit transformation, and video workflow applies to the whole group at once.

#Models

Photobooth ships a curated mix of fast SDXL turbo / lightning models, full-quality diffusion models, identity-preserving edit models, and the OpenAI frontier model:

All models run on the Sogni Supernet — you don't manage them, install them, or pay per call. Logged-in users hit the SDK directly for lower latency; demo users go through the same backend used by every other Sogni app.

#Choose Your Video Style

Photobooth has the broadest video toolkit of any Sogni superapp. The Choose Your Video Style picker exposes seven distinct workflows, all driven off a single source portrait. Two model families (Wan 2.2 14B and LTX-2.3 22B) split between them, with four quality tiers each — Fast (≈4 steps), Balanced (≈8 steps), Quality (≈20 steps), Pro (≈30 steps). Default output is 480p at 32 fps (Wan's 16 fps interpolated post-render); LTX-2.3 generates natively at 24 fps and supports longer clips.

#See them in action

#Audio for Sound to Video

You can upload an existing audio clip or generate one on the fly with ACE-Step 1.5 (Turbo or SFT) — Photobooth's built-in music generator supports 40+ languages, BPM/key/scale control, and clips from 10 seconds to 10 minutes.

#AI editing & transformations

The Image Edit Styles category bundles 19 dedicated transformation presets that go beyond restyling — they reshape the subject:

Lego · claymation · Pixar · Simpsons · Mario Kart · Minecraft · Fortnite · pixel art · pop art · Ukiyo-e · doodle art · tattoo flash · neon · Viking · boxer · bodybuilder · bobblehead · cloned subject · "hats and glasses"

Each preset is wired to a Qwen Image Edit (or Flux.2 Dev / GPT Image 2) backbone so likeness is preserved through the transformation. You can also use them in combination with any other workflow — turn yourself into a Lego version, then animate it into a video; or run the Viking through camera-angle remapping for a four-pose character sheet.

#Camera-angle remapping

Photobooth's Multiple Angles workflow re-renders the same subject from any combination of:

• 8 azimuths — front, front-right, right, back-right, back, back-left, left, front-left
• 4 elevations — low-angle, eye-level, elevated, high-angle (bird's eye)
• 3 distances — close-up, medium, wide

That's 96 unique angle combinations off a single source portrait — plus six one-tap presets (3/4 portrait, profile, hero shot, overhead, close-up, over-shoulder) and four multi-angle bundles (Simple Zoom Out, Zoom Out 360, Zoom Montage, Portrait Trio) that fill a character sheet in a single batch. Maximum 16 angles per request. Backed by the Multiple Angles LoRA on Qwen Image Edit 2511 Lightning.

The standalone Sogni 360 superapp is a focused expression of this same engine if you only need the camera-angle work.

#Workflow recipes

Professional headshot. Upload a clean portrait → Pro / Editorial category → 3:4 aspect → 16-image batch on JuggernautXL or RealVisXL → pick the best → run through Sogni Restore for a final upscale.

Character portrait. Fantasy & Sci-Fi or Anime category → 4-image batch → run the winner through the Portrait Trio multi-angle preset to build a consistent character sheet.

Custom theme for an event. Open Personalize, type the theme ("90s prom"), generate the prompt series, share the event link — every attendee can use the same custom vibes on their selfies.

Music video clip. Generate a portrait → Image-to-Video → Wan 2.2 Quality at 720p → optionally add audio via Sound-to-Video for a 5–8 s shareable clip.

Lip-synced short. Upload a song clip with your selfie → Sound to Video on Wan 2.2 S2V (or LTX-2.3 IA2V for longer clips) → beat-matched 5–10 s video.

Dance Montage. Pick your portrait → Motion Transfer with a reference dance clip → generate a backing track in ACE-Step → Photobooth stitches it into a multi-style dance reel where you're the star.

Group photobooth at an event. Auto-detected group input routes to Flux.2 Dev / Qwen Image Edit → batch generate across a curated event vibe → kiosk mode prints a QR-watermarked card on the spot.

#Aspect ratios

Seven supported ratios so output is right-sized for the destination:

Mobile (4:7) · 35 mm (2:3) · Portrait (3:4) · Square (1:1) · Landscape (4:3) · Cinema (3:2) · Widescreen (16:9)

Pick before you batch — re-rendering after the fact is cheap, but cropping a portrait into a widescreen rarely looks right.

#Sharing, kiosks, and prints

• QR watermark — every output can be tagged with a QR linking back to a sharable gallery; configurable position and size for kiosk deployments.
• Kiosk mode — full-screen UX designed for event laptops, tablets, and standing kiosks.
• Event branding — custom themes via Personalize stay scoped to the event so all attendees share the same vibe pool.
• Physical print fulfilment — order prints of any generation directly from the app; Photobooth integrates with Stripe and the Sogni print pipeline.

#Tips for best results

• Lighting beats resolution. Even lighting on the face matters more than camera megapixels.
• One face per shot for portrait modes. Use Group Photos mode (or let the auto-detect catch it) when there's more than one person.
• Pick an aspect ratio before generating. 3:4 for portraits, 1:1 for avatars, 16:9 for video.
• Use Fast mode for exploration, Pro for the keeper. Burn batches in Fast, then re-render the winner at higher quality.
• Try Personalize on niche themes. "World famous paintings" is a fine prompt; "Dutch still-life paintings from the 1600s with hidden symbolism" gets you something nobody else has.
• Multi-angle remapping is the fastest character sheet. A single good portrait + Portrait Trio = a full reference sheet in one batch.
• Save winners to Favorites. The Favorites category persists across sessions and is the easiest way to build your personal go-to set.

#See also

• Sogni Photobooth — Real-life experiences — kiosk, event, and hardware deployments.
• Sogni Client SDK — the SDK that powers Photobooth.
• Sogni 360 — the camera-angle remap engine as a focused standalone app.
• Sogni Supernet — the network running every generation.