Create a high-fidelity desktop application UI for “Salvor AI”, an AI identity-to...

Create a high-fidelity desktop application UI for “Salvor AI”, an AI identity-to-intent creation studio. Use the supplied original Standard UI image as the PRIMARY structural reference and the other supplied images as secondary visual references. The product is based around: IDENTITY → INTENT → GENERATE Users upload photographs to create reusable identities of themselves, friends, family, pets, characters, or other subjects. They then use those identities to create intents such as placing them in different locations or eras, changing clothing, hairstyle, hair color, tattoos, accessories, actions, poses, environments, or visual styles, and can ultimately generate images or video. The interface should NOT require users to write prompts. The controls themselves establish the intent. IMPORTANT BALANCE: The UI should contain meaningful imagery, but it must NOT become an image gallery or a wall of visual cards. Use images strategically: Identity thumbnails A few visual intent presets Small visual references for selected options The generated output Most of the interface should consist of clean controls, labels, selectors, buttons, compact cards, tabs, and structured sections. Aim for roughly 60% interface/control structure and 40% imagery/visual content, rather than making the screen primarily photographic. VISUAL LANGUAGE: Premium dark navy/charcoal desktop application, subtle panel borders, restrained purple/violet accents, occasional blue highlights, modern typography, rounded cards, soft shadows, excellent spacing, cinematic but restrained imagery. It should feel creative, polished, comfortable, and approachable rather than technical. NO upgrade buttons, pricing, advertisements, subscription messaging, sales content, or promotional elements. LAYOUT: TOP HEADER: A clean application header containing: Salvor AI logo/name Current creation/project A few understated utility controls Clearly visible Advanced toggle The Advanced toggle should currently be OFF. LEFT SIDEBAR — IDENTITIES: A compact vertical sidebar called “Identities”. Show a small number of identity thumbnails with names, for example: Me Friend Dog Character Use modest portrait thumbnails, approximately thumbnail/card size rather than large photographic panels. At the top provide a clear: “+ Create Identity” When creating an identity, show a simple upload interface: “Upload photos to create an identity” with Upload Photos / Take Photo / Drag & Drop. The identity creation experience should make it obvious that identities are reusable assets, not one-time image uploads. MAIN WORKSPACE: SECTION 1 — IDENTITY Show the selected identity as a compact horizontal card: [thumbnail] Me Identity ready Allow “Change” and “Add Identity”. Multiple identities should be possible, but remain simple. SECTION 2 — INTENT Use a clean heading: “What do you want to create?” Provide a balanced mixture of a few visual preset cards and structured controls. PRESET INTENTS: Show approximately 6–8 small visual presets, each with a thumbnail and concise label: Renaissance Portrait Vacation Historical Street Style Fantasy Professional Adventure Custom These should be small enough that they function as quick starting points, not as a gallery. Selecting a preset establishes sensible intent settings automatically. Below the presets, provide the primary intent controls as structured rows or compact panels: WHERE Location: Florence / Tokyo / New York / Beach / Mountains / Custom WHEN Era: Modern / 1920s / Renaissance / Medieval / Future / Custom LOOK Clothing: Current / Formal / Medieval / Streetwear / Custom APPEARANCE Hair / Tattoos / Accessories / Other DOING Portrait / Walking / Sitting / Action / Interaction / Custom STYLE Photorealistic / Cinematic / Vintage / Fantasy / Editorial OUTPUT Image / Video These controls should primarily be pills, dropdowns, segmented controls, compact selectors, and small thumbnail-supported choices. Do not make every option a giant image card. For options where imagery is genuinely useful, use a small thumbnail alongside the label. CONTEXTUAL ADVANCED CONTROLS: Every major category should have a subtle settings/sliders icon. Clicking it opens a compact flyout. Examples: Hair → style, length, color Appearance → tattoos, makeup, facial hair Clothing → material, color, accessories Pose → detailed pose and body positioning Location → environment, weather, time Camera → framing, lens, depth of field Identity → preservation strength and feature locking These flyouts are important because Standard mode should remain simple while still allowing precision when requested. RIGHT SIDE — OUTPUT: Provide a meaningful but contained output panel. Before generation, show a modest preview/summary of the selected intent, combining the selected identity with the chosen intent. For example: Me Renaissance Royal clothing Portrait Cinematic Then show a clean Generate button. After generation, the panel becomes the generated image/video preview. The output image should be substantial enough to feel satisfying, but should NOT occupy almost the entire screen. Under it provide a small set of actions: Variations Change Intent Remix Create Video Save Share SOCIAL: Include a compact “Create for Social” option with small destination buttons: Instagram TikTok YouTube Pinterest Profile Do not make social generation a large section. OVERALL UX: The user should be able to understand the entire workflow without reading instructions: Pick or create an identity. Choose a preset or establish an intent. Adjust a few simple attributes. Generate. The screen should contain enough imagery to feel visually creative and inspiring, but enough interface structure that it clearly reads as a powerful creative application rather than an image browsing experience. STANDARD MODE SHOULD FEEL: Simple on the surface. Visual where useful. Structured everywhere else. Powerful underneath. ADVANCED TOGGLE: The Advanced switch must be clearly visible in the header. When turned ON, the user stays in exactly the same creation and the same UI. Advanced mode should expose additional controls within this same layout rather than opening an entirely different application. The Standard UI should therefore be designed specifically so that controls can expand into the Advanced UI. High-fidelity production-ready desktop SaaS interface, 16:9 screenshot, realistic UI, sophisticated dark creative software, balanced imagery and controls, restrained visual hierarchy, no browser chrome.