Sogni: Learn logo
Markdown

Sogni Create for Web and Android

Sogni Create for Web is the shared codebase behind app.sogni.ai and the Android app on Google Play. Sign in from any modern desktop or mobile browser, or install the Android app for a Play Store-managed mobile shell around the same Sogni Create workspace. Either way, you get the same image, video, and music toolkit backed by the Sogni Supernet — no local GPU required.

Forgot your username or password? Follow the Web & Android sign-in recovery guide to find saved credentials in Chrome, Safari, Edge, Firefox, and other password managers on desktop or mobile.

#Launch Sogni Create on Web →

#Get Sogni Create on Google Play →

Sogni Create on desktop and mobile browsers

#What it does

  • Cross-platform. Runs on Windows, macOS, Linux, ChromeOS, iPad, and Android browsers, plus the Android app on Google Play. Same UI, same account, same gallery.
  • Zero-install web, store-installed Android. Open a browser tab when you want no install, or install from Google Play when you want launcher access and app-store updates on Android.
  • One account everywhere. Your prompts, gallery, balance, and Spark Points sync with Pocket and Studio — start a render in the browser or Android app, finish it on your phone or Mac.
  • Powered by the Sogni Supernet. Jobs are dispatched to a distributed worker network, so a low-end laptop renders as fast as a workstation.

#Features

The web and Android app share most of their feature set with Studio, so behavior is documented once there and the same behavior applies here. The main menu holds Home, Create, Video, Music, Speech, Mixer (the Clip Mixer), Creations (your recent projects), Wallet (Account in the Android app), Usage, Rewards, Settings, Community, and Help.

  • Image generation (Create) — text-to-image, image-to-image, batches, dynamic prompts, and an AI Prompt Creator that drafts prompts for the currently selected model.
  • Video generation (Video) — Sogni-native and premium hosted model families exposed from the Video home page:
    • LTX-2.3 — Text→Video, Image→Video, and Sound→Video. Up to 4K, 60 fps, with native audio generation.
    • WAN 2.2 — Text→Video, Image→Video, Motion Transfer (drive a character from a reference video) and Replace Subject (keep the scene, swap the subject). Up to 1,048,576 pixels per frame (1024×1024, or 1280×720 widescreen), 32 fps.
    • Seedance 2.0 — premium hosted full and Mini tiers for text, image, audio, and video-reference workflows. Use the full tier when a 4K render is required. Both accept photos and clips of real people after a one-time likeness agreement.
    • Seedance 2.5 — latest premium hosted Seedance tier and current storyboard default: 4–30 second clips with an expanded multimodal reference budget at 480p, 720p, or 1080p (no 4K). It accepts photos and clips of real people after a one-time likeness agreement.
    • Seedance 2.5 Uncensored — Seedance 2.5 on a more permissive content policy, with every 2.5 capability at a higher price. It appears next to Seedance 2.5 only while the Sensitive Content Filter is off, and asks for a one-time likeness and consent agreement before the first render.
    • Seedance 2.0 Mini Uncensored — Seedance 2.0 Mini on a more permissive content policy: 4–15 second clips at 480p or 720p, at a higher price than Mini. Like Seedance 2.5 Uncensored, it appears only while the Sensitive Content Filter is off and uses the same one-time likeness and consent agreement.
    • HappyHorse 1.1 — premium hosted text-to-video, image-to-video, and image-reference video with native synchronized audio.
    • MiniMax H3 — Standard, Balanced, LightX2V Turbo, and FastH3 Turbo for text-to-video, image-to-video, and first/last-frame video, plus Reference-to-Video (including FastH3 Reference) and FastH3 Sound to Video, with synchronized stereo audio. The Keyframes block pins up to eight more images at exact moments inside an Image to Video, Sound to Video, or Reference to Video clip, so you can storyboard it shot by shot; see Intermediate keyframes. Continue this video… › Seamless in a video's menu, or a Source video on Text to Video, picks the clip up where it ends and carries its motion and sound on; see Extend a video.
  • Music generation (Music) — full songs with AI-written lyrics, instrumentals, or short loops, powered by MiniMax Music 3 (up to 300 seconds, a ceiling the song may end before) and the ACE-Step model family (10–600 seconds, with controls for BPM, key, scale, and time signature), plus a curated library of musical-style presets. See Creating audio.
  • Clip Mixer — open Mixer in the main menu (Clip Mixer on phones) or go to app.sogni.ai/clip-mixer to trim, split and arrange your clips and images into one video, with transitions and music, then export it as an MP4. A video's menu in Creations also has Add to Clip Mixer. Your browser builds the video on your own device, and each browser keeps its own Clip Mixer project. See AI transitions in Clip Mixer and Clip Mixer shortcuts. If an export won't move past 0%, see Clip Mixer export stuck at 0%.
  • ControlNet — pose, depth, canny, and other guidance maps driven by reference images for ControlNet-capable models.
  • Qwen Image Edit — when a qwen_image_edit_* model is selected, the Create tab exposes Qwen reference-image controls for instruction-style image editing.
  • Seedream 5.0 Pro Uncensored — ByteDance's Seedream 5.0 Pro on a more permissive content policy, for text-to-image and edits from up to ten reference images, priced per image in Premium Spark. It appears in the image model picker only while the Sensitive Content Filter is off, and asks for the same one-time likeness and consent agreement as the Seedance models before the first render.
  • FLUX 3 Image Klein — one model for text-to-image and edits from up to ten reference images. Pick it under Image Editing & Mixing in the model picker, or choose it from Image Editing on Home to start an edit.
  • LoRAs and My LoRAs — on Krea 2, Qwen Image Edit 2511, and MiniMax H3 models, open LoRAs → Browse LoRAs (Manage LoRAs once some are attached) to stack up to 8 LoRAs; the order matters. Unlimited members can also open My LoRAs (Beta), paste a public Hugging Face .safetensors file link or a Civitai model or version link, name it, confirm permission to use it, and Import. Imported LoRAs show only while the Sensitive Content Filter is off. See Personal LoRAs for file requirements and limits.
  • Image enhancement — per-result enhance menus add Light / Medium / Heavy passes with Krea 2 Turbo, Z-Image Turbo or FLUX.1 [schnell], plus camera-angle / perspective tools.
  • Model picker — browse the full Sogni model catalog (Krea 2 Turbo, FLUX 3 Image Klein, SD 1.x / 2.x, SDXL, SD3, Flux, Z-Image, Qwen Image Edit, WAN, LTX-2.3, Seedance 2.0, Seedance 2.5, Seedance 2.5 Uncensored, Seedance 2.0 Mini Uncensored, and HappyHorse 1.1) and switch with a click. See Models.
  • Generation settings — steps, guidance, scheduler, seed, aspect ratio, and other generation controls map 1:1 with Studio.
  • Recent Projects, Wallet, Rewards — full project history, balance and spend tracking, and the Rewards / Spark Points program in the same tab strip.

Surfaces that exist in Studio but not on web or Android include the in-app Chat assistant, on-device CoreML rendering, .sogni project files, the Activity window, and Worker Mode. Use Studio or Studio Pro if you need those.

#Clip Mixer

#AI transitions in Clip Mixer

Tap the marker between two clips. The AI Transition tab opens first. Sogni generates a short clip that carries the end of the first clip into the start of the next.

  • Style. Auto writes the transition from your two frames. You can also pick a style, or describe your own under Custom.
  • Model. MiniMax H3, the default, makes transitions of 5 to 15 seconds and can pass through key images. LTX-2.3 makes transitions of 2 to 4 seconds.
  • Key images. With MiniMax H3, add up to eight pictures under Key images, each at its own moment, and the transition passes through each one. A picture with a different framing, place or light is cut to rather than blended.
  • Takes. Regenerate adds a new take and keeps the one you have. Open the transition to preview its takes, use the one you like, or delete the ones you don't want.

Turn into video, in a still's menu, makes a video on MiniMax H3. When you add an image, Subtle motion starts on MiniMax H3 (5 to 15 seconds), or pick LTX-2.3 for 2 to 8 seconds.

#Clip Mixer shortcuts

On a computer, the Clip Mixer responds to these keys while no menu or window is open:

Key What it does
Space Play or pause
← / → Step one frame back or forward (with Shift, one second)
↑ / ↓ Go to the previous or next clip
Home / End Go to the start or the end
S Split the clip under the playhead
⌘Z / Ctrl+Z Undo the last edit
⇧⌘Z / Ctrl+Y Redo it

Split at playhead and Duplicate are also in each clip's menu, and the arrows at the top of Clip Mixer undo and redo.

#Workflows

Quick start. Sign in, type a prompt, hit Imagine. The default model and settings are tuned for fast first-try results.

From phone to desktop. Start a generation on mobile during a commute, open the same gallery on your desktop browser later, and enhance or remix without re-uploading anything.

Shared sessions. Because everything lives in your Sogni account, switching between the web app, the Android app, Pocket, and Studio mid-project just works — no export step.

#Install options

app.sogni.ai is a Progressive Web App. On supported browsers you'll see an Install prompt the first time you visit — accept it (or use your browser's Install app / Add to Home Screen menu) to get:

  • A standalone window with the Sogni Create icon, no browser chrome.
  • An icon on your desktop dock, app launcher, or phone home screen.
  • A service-worker-backed offline page when the network drops.

On Android, the Google Play app is the recommended install path when you want a dedicated app entry, store updates, and a mobile app shell around the same Sogni Create web experience. The PWA remains useful when you want Sogni Create without a store install.

#Tips

  • Use Chrome, Safari, Edge, or Firefox. WebSocket support and modern JS are required; very old browsers won't connect to the Supernet.
  • Mobile is fully featured. The shared app reflows to a touch UI on phones with its own mobile toolbar.
  • Install the Android app or PWA (see above) when you want it to feel like a dedicated app instead of a browser tab.
  • Need Worker Mode or on-device rendering? That's Studio Pro — web and Android render on the Supernet.

#Troubleshooting

#Clip Mixer export stuck at 0%

The Clip Mixer builds your video in your browser, not on Sogni's servers, so its progress normally starts moving within seconds. If it stays at 0% ("Preparing export…") for a minute, the export has stalled and waiting longer usually won't help. The export window then says This export seems stuck and offers Reload page.

  1. Reload the page and export again. Your Clip Mixer project is kept. Keep the tab open in front and your device awake until the export finishes.
  2. If it stalls again, try another up-to-date browser on a computer, such as Chrome or Edge. Each browser keeps its own Clip Mixer project, so add your clips there again: drag the video files you downloaded onto the page, or use Add video or Batch import. Videos from the last two days can also be sent from Creations with Add to Clip Mixer. Sogni keeps generated media for about two days, so download clips you want to keep.
  3. Until you have your finished video, don't clear the browser's site data, don't use Clear in the Clip Mixer, and choose Keep my data if you sign out. For clips you didn't download that are more than two days old, that browser may hold the only copy.

If it still won't export, contact support with your browser and its version, your device and its operating system version, and whether the export window says "Preparing export…" or "Rendering frames…".

#See also

Last updated 2026-10-22