Travel diary collageExample 01gpt-image-2
gpt-image-2 is available through TTAPI. OpenAI image generation through TTAPI, with fixed-price GPT Image models and their corresponding official-transfer routes.
Travel diary collageExample 01
Cool-white tennis portraitExample 02
Impasto fashion portraitExample 03
Summer through iceExample 04Prompt to image, in context
[Task] Transform the person in 【Character Reference Image】 into a “nighttime oiran-style photoshoot in a Japanese tatami room” aesthetic: an indoor tatami room setting, with a strong on-camera direct flash creating a high-contrast, slightly dirty/gritty nighttime snapshot texture. The subject wears a red-and-black brocade kimono / oiran-style ornate outfit, holds a black-and-red folding fan partially covering the face, with an opened paper umbrella behind them. Outside the wooden window is a nighttime garden with hanging strands of white flowers. The character’s identity must remain consistent (facial features and face shape unchanged). [Identity Consistency] * Strictly preserve 【Character Reference Image】: face shape contour, facial feature proportions (brows/eyes / nose bridge / lips), sense of age, gender, and skin tone base * Allowed: makeup, hairstyle, clothing, pose, props, environment, and lighting reconstruction * Forbidden: face swapping, generic influencer-face beautification, anime stylization, excessive smoothing / plastic skin [Aspect Ratio & Composition] * Vertical 2:3 * Framing: seated half-body to two-thirds body shot (including legs and the spread of the garment hem), with the subject occupying the lower-middle area of the frame * Composition: subject positioned slightly below center; wooden window lattice clearly visible in the background; an opened paper umbrella appears in the rear-left area as a compositional counterweight * Camera angle: slightly high angle with a gentle downward tilt, carrying a casual “handheld night snapshot” realism, not a formal studio setup [Pose & Body Language] * Pose: kneeling / side-seated on tatami (hips on the floor, legs bent and extending outward to one side), with the upper body leaning slightly forward * Shoulders / neck: exposed shoulder or a slipped collar, visible collarbones, with a subtle sensual undertone * Hands: one hand raises a black-and-red folding fan, covering a small portion of the face; the other hand supports the fan or gathers the sleeve * Head & gaze: head slightly lowered, eyes glancing toward the camera from the edge of the fan or half-lidded (seductive, dangerous, provocative, but restrained) * Expression: no smile; cold beauty with a teasing undertone [Makeup — sensual oiran / geisha-inspired makeup] * Base makeup: pale, cool-toned powdered complexion (while retaining real skin texture, no plastic skin smoothing) * Eye makeup: red eyeshadow / flushed red tones spreading across the upper eyelids and lower eyelids, elongated eyeliner, creating an overall “drunken red / yandere-like” atmosphere * Lip makeup: deep red or dark crimson, with sharply defined lip shape * Optional: slight facial blush, red shading under the eyes (to intensify the eerie sensuality) [Hairstyle & Hair Accessories] * Hairstyle: high updo / large traditional coiffure * Hair accessories: golden tassel hairpins + a large red flower on the side (must be clearly visible), with a small amount of dangling ornaments * A few loose strands against the face are allowed to enhance the realism of a nighttime snapshot [Clothing & Material] * Main outfit: a red-and-black brocade kimono / ornate oiran robe, black outer layer with gold-and-red floral patterns, bright glossy red inner layer * Material: high-sheen satin / brocade, with strong reflective highlights along the folds * Pattern: intricate floral motifs / gold-leaf-like ornamental patterning (avoid flat printed-texture appearance) * Extra fabric: A piece of blue brocade (blue background with intricate patterns) appears in the foreground, serving as a color contrast (can be draped over the legs) [Props & Setting] * Folding fan: black fan leaf + red fan ribs / red edging + subtle cherry blossom / floral decoration * Paper umbrella: an opened oil-paper umbrella placed behind the subject on the left side (may be partially in frame) * Environment: tatami flooring; wooden walls and wooden window lattice; nighttime garden outside the window, with tree-branch shadows * Important accent: a hanging strand of white flowers by the window (like wisteria / cascading white blossoms), forming a vertical visual focal point [Lighting & Color] * Lighting: must be “on-camera direct flash” * Bright highlights on the subject’s skin, hard shadows, relatively dark background * Slight blown-out edges and reflective hotspots (especially strong reflections on the kimono’s satin surface) * Color: strong red-black contrast, with a touch of vintage noise and dirty nighttime texture; no cinematic soft studio lighting * Texture: slight grain / noise, mild compression feel, authentic nighttime snapshot atmosphere [Image Quality & Style Direction] * Realistic photography, like a real nighttime tatami-room portrait shoot (not CG render, not anime) * The background should not be overly blurred into a studio-like glamour shot; preserve the authentic spatial feel of a casual snapshot [Negative Constraints] * Do not include: front-facing ID-photo composition, softbox studio lighting, cinematic dreamy diffusion glow * Do not include: modern street scenes / neon city backgrounds; no palace-scale grand environment * Do not include: fresh ethereal ancient-Chinese fantasy vibe, celestial/xianxia styling; no sweet smiling expression * Do not include: cheap cosplay fabric, flat printed patterns, plastic skin, excessive skin smoothing * Do not include: extra fingers, malformed hands, facial distortion, unfocused eyes [Output] Generate 1 image: high identity consistency with 【Character Reference Image】; pose, fan partially covering the face, paper umbrella, tatami room wooden window, night scenery outside with white flower strands, red-black brocade texture, and the hard direct-flash nighttime atmosphere must all be fully realized.
Cinematic portrait photography, ultra-photorealistic, 2160x3840 vertical composition, 50mm or 85mm portrait lens rendering, shallow depth of field, clean translucent summer natural-light color grading — not overly yellow, not over-filtered. Subject: a young beautiful adult East Asian woman, [describe face shape and features, e.g. soft heart-shaped face, refined classical features, bright almond/fox eyes, petite nose bridge, naturally full lips], overall vibe sweet, sunny, energetic, cute with a touch of allure. Gaze highly engaging — bright, clear, natural catchlights, as if it speaks; corners of mouth slightly lifted, expression gentle, vivid, natural. She walks along a [scene, e.g. garden stone path / tree-lined lane / European street / courtyard], right hand reaching back to hold the hand of someone behind her; only their hand appears in the lower-left corner — like a first-person couple's POV snapshot. She glances back at the camera while her body stays in a forward walking motion, posture elegant and natural, clearly a candid captured moment with a faint in-love feeling. Long [hair color] hair, [style, e.g. naturally wavy / relaxed big waves / airy bangs / half-up], many strands tousled and flying in the wind, richly layered and dynamic. Strong natural side-backlight rims the hair edges — clean, crisp rim light and semi-translucent glow, hair edges lit as if by sunlight, light and luminous. This is the core highlight of the image. She wears [outfit, e.g. white lace slip dress / beige slip dress / light-blue short-sleeve top with white skirt / light-pink fitted dress], fabric texture natural, material light and soft. Bright natural summer sunlight realistically warms her skin, shoulders, collarbone, and clothing with soft, clean highlight transitions. Skin texture: extremely realistic — visible fine pores, natural skin texture, faint imperfections, subtle tone variation, soft sheen. Cheeks, nose tip, shoulders show natural delicate gradations in sunlight. Translucent, healthy, real and refined — no plastic look, no waxwork, no over-smoothing. Background: soft atmospheric blur, never distracting. Avoid: over-smoothing, plastic skin, CG look, anime look, wig look, stiff expression, dead eyes, stiff poses, overall yellow cast, overexposed face, distorted features, wrong fingers, deformed hands, cluttered background, heavy influencer retouching.
Create a serene premium matcha advertising poster for a fictional brand "KYO GREEN". FORMAT: 4:5 vertical, elegant wellness layout. SUBJECT: a ceremonial bowl of vivid green matcha with delicate foam, a bamboo whisk mid-motion, matcha powder gently dusting down, fresh green leaves. BACKGROUND: calm soft-green to cream gradient with rice-paper texture. LIGHTING: soft diffused natural light, delicate highlights. TYPOGRAPHY: refined minimal logo "KYO GREEN" top-center, tiny tagline "Pure ritual" bottom. FINISH: 8k, hyper-realistic, tranquil finish.
Produce an ultra-realistic commercial beverage poster for KESAR Almond Saffron. Treat the uploaded product or subject image as the exact packaging reference; preserve label typography, can geometry, color, and brand placement. Center one chilled can at a slight dynamic angle, covered in crisp condensation, surrounded by a controlled burst of the real flavor ingredients. Use a premium #F4A300 background with layered light, subtle atmospheric particles, and a reflective surface that grounds the product. Render the exact headline "Rich by nature" and supporting line "A chilled blend with roasted almond and saffron." with a clear hierarchy that does not overlap the can. Make the product look cold, tactile, and immediately appetizing. Output one 4:5 poster with no fake badges, no invented ingredients, no watermark, and no duplicate cans.
Hyper-realistic premium product advertisement: an oversized futuristic comfort clog sits on a smooth glossy reflective floor. A modern model in soft neutral-toned athleisure (off-white / beige) leans casually against the giant shoe with a relaxed, confident posture. Backdrop: clean gradient flowing from soft sky blue into subtle lavender, with massive bold sans-serif typography reading "STEP INTO EASE" stretched vertically, partially tucked behind the subject. Lighting: high-end studio lighting, soft highlights, gentle floor reflections, and a subtle rim light tracing the model and the product for depth. Composition: editorial magazine layout, subject perfectly centered, generous negative space, luxury campaign mood. Small minimal copy at the bottom: "Designed for all-day comfort. Made to move with you." Style: ultra-clean Apple-style minimalism crossed with a fashion campaign, hyper-realistic, premium commercial photography, 8K, razor-sharp detail.
Hyperrealistic full-length promotional photo of a giant mango ice cream in a waffle cone, immersed in a swirl of glossy golden mango cream, top of the ice cream decorated with oversized mango cubes. Mango chunks swirl chaotically in creamy storm. Ice cream appears falling with glossy textures, background: Santorini with white houses and blue domes, bright summer light. Shot on Camera Sony A9. --ar 2:3 --stylize 300
{ "style_name": "Neon Liquid Interface Dossier", "style_slug": "neon-liquid-interface-dossier-style", "style_version": "2.1", "canvas": "9:16 or 16:9 — recomposed natively for each, never mechanically cropped", "color_palette": { "carbon_black": "#090b0d", "ink_graphite": "#24292e", "silver_gray": "#a8adb0", "paper_white": "#edf0ef", "molten_yellow": "#ffd52f", "signal_orange": "#ff6b22", "hot_coral": "#ff3e56", "electric_pink": "#f32a8a", "spectral_cyan": "#22d4d8", "deep_teal": "#075c66" }, "color_behavior": "~85% black / white / graphite / silver; ~15% saturated molten neon concentrated into ONE continuous fluid or membrane with concentric warm-to-cool transitions, glossy speculars and dark pooled edges", "fidelity_anchors": [ "One monumental monochrome hero dominates the center, cropped by the frame, near-symmetrical frontal silhouette", "Photographic anatomy fused with smooth reflective liquid-metal shells — surreal but materially believable", "Oversized wide geometric uppercase type spans the frame, partially obscured by the subject, behaving as architecture not caption", "Two or three circular inspection lenses magnify metallic or halftone detail at different scales; at least one crops against a frame edge", "Thin white or charcoal interface lines use right-angle routes with endpoint dots, bracket frames, bar graphs, compact status panels", "High-contrast monochrome base: crushed blacks, blown pale highlights, coarse photocopy grain, halftone dots, weathered edges", "Shallow poster-like depth despite glossy 3D materials — hierarchy from overlap and scale, not scenic perspective", "Hard frontal sculptural light: silver rim reflections, dark cavities, a luminous neon core, no cinematic haze", "Mood is clinical, uncanny, editorial — an experimental technology dossier printed through an imperfect analog process" ], "typography": { "display": "extra-wide ultra-bold squared geometric uppercase, tight leading, slight photocopy bloom; ONE invented word, split across the top and lower thirds", "microcopy": "narrow monospaced uppercase, small, generously tracked, short lines, clipped fictional diagnostic phrasing", "hierarchy": "display word → neon material event → hero → lenses → microcopy", "text_rule": "all wording original, brief, unbranded, non-scannable" }, "composition": [ "Build around one central hero and one dominant material event", "Reserve the outer fifths for narrow data columns and short fictional labels", "Place lenses at distinctly different scales; crop one against an edge", "Route connectors in crisp 90° paths ending in small solid nodes", "Keep the neon compact and continuous — never rainbow accents across the frame" ], "avoid": [ "color spread across the whole image", "game character sheet, movie poster, software dashboard, generic vaporwave", "pastel gradients, cinematic fog, warm natural light, painterly or cartoon rendering", "logos, watermarks, usernames, platform UI, QR or scannable patterns", "repeating the same silhouette, lens placement, or connector routing between cases" ], "swap_variables": { "specialist": "Specialist", "main_text": "Main Text", "color_object": "Color Object", "measurement_ladder": "Measurement Ladder", "lens_subjects": "Lens Subjects", "accent_glyph": "Accent Glyph" } }
Use the uploaded reference image as the exact identity base for the main subject. Preserve their authentic facial structure, recognizable appearance, hairstyle, body proportions, skin texture, and overall identity with high consistency. Completely ignore any unrelated background from the original image. Create a hyper-chaotic early-2000s Japanese digicam snapshot aesthetic with raw paparazzi energy and accidental comedy. Scene: the subject is sprinting wildly through a busy city street while yelling and laughing in pure chaos, desperately trying to catch a mischievous cat that just stole a large shiny fish from a market stall. The cat is gripping the slippery silver fish tightly in its mouth while running directly toward the camera. Composition must prioritize TWO focal subjects: 1. PRIMARY FOCUS → the cat's face 2. SECONDARY FOCUS → the human subject's face The cat dominates the foreground frame with its face pushed absurdly close into the lens, partially cropped near the lower-right side. Massive fisheye distortion stretches its features dramatically - huge bulging eyes, exaggerated whiskers, detailed fur texture, fish flapping violently with droplets and motion streaks flying outward. Behind the cat, the subject is charging forward aggressively with intense energy, reaching toward the animal mid-run. Their expression is loud, chaotic, and comedic, eyes locked directly onto the cat to create strong interaction and pursuit tension. Styling: Randomized layered Y2K Tokyo streetwear inspired by chaotic Shibuya nightlife fashion - oversized zip hoodies, vintage mesh layers, mini bags, low-rise pieces, striped arm warmers, chunky accessories, early-2000s sneaker styling, messy fashionable youth aesthetic. Camera & framing: extreme low-angle fisheye lens, ultra-wide invasive perspective, severe Dutch angle, asymmetrical framing, warped spatial distortion, exaggerated depth compression, aggressive foreshortening. Motion treatment: insanely intense movement blur with radial streaking, shutter drag, CCD sensor smearing, directional speed trails, accidental camera shake, stretched highlights, warped moving objects, imperfect chaotic snapshot timing. Environment: sunny crowded urban street with pigeons exploding into flight across the frame, scattered newspapers and debris flying through the air, warped pedestrians, market elements smeared by motion, overexposed sunlight leaking into one corner of the frame. Lighting & texture: harsh direct flash combined with bright outdoor sunlight, dirty CCD digicam texture, blown highlights, chromatic aberration, digital noise, bloom artifacts, compressed early-2000s camera look, gritty nostalgic cyber-chaos atmosphere. Ultra-raw candid energy, messy composition, humorous accidental masterpiece aesthetic.
{ "style_rules": { "background": "white", "typography": "black condensed bold, distressed texture, top-left, 2-3 word imperative", "photo": "full-color action shot, subject airborne, fragmented into overlapping tilted panels", "ink_brush": "black calligraphic brush arcs sweeping around subject", "wireframe": "thin black architectural line drawings (scaffolding) filling negative space", "speed_lines": "white diagonal slash lines for motion", "metadata": "bottom-left, small caps, 3-line descriptor", "format": "16:9 horizontal" }, "negative": "black background, serif fonts, motion blur, clean minimal layout, gradients", "variables": { "subject": "", "headline": "", "metadata": "" } }
What you can build with gpt-image-2
Prompt driven generation
Submit a concise prompt and get production-ready image output through one TTAPI job.
Core controls only
Keep integration simple with prompt, model, output settings, and callback handling.
Async result flow
Use polling or webhook callbacks so long-running generations do not block your UI.
Ready for product UI
Return OpenAI Image results that can be displayed, stored, or passed into downstream workflows.
Compare fixed and official channels
Choose predictable fixed pricing or the official-transfer route for generation and editing.
Fixed-price non-official channel.
Pricing sourceOfficial OpenAI image generation and editing are billed at 20% off official pricing.
Pricing sourceRequests by channel
Switch between non-official and official-transfer requests. Each channel keeps its own endpoint, fields, and documentation.
Generate
Submit an asynchronous fixed-price GPT Image task and retrieve the result by job ID.
Headers
TT-API-KEYstringrequiredYour TTAPI API key.
Content-TypestringrequiredUse application/json.
Body
promptstringrequiredText description of the image.
modelenumoptionalgpt-image-1.5, gpt-image-2, or gpt-image-2-plus; defaults to gpt-image-2.
referImagesstring[]optionalReference image URLs for image-guided generation.
aspect_ratioenumconditionalAspect ratio control supported by gpt-image-2.
sizestringconditionalWIDTHxHEIGHT output size supported by gpt-image-2-plus.
hookUrlstringoptionalCallback URL for asynchronous completion.
Integrate the selected channel
Choose a channel and operation, then copy the matching endpoint, request body, and code sample.
Generate integration
POST https://api.ttapi.io/openai/gpt/generations
curl --request POST \
--url https://api.ttapi.io/openai/gpt/generations \
--header 'TT-API-KEY: $TTAPI_KEY' \
--header 'Content-Type: application/json' \
--data '{"prompt":"minimal product photograph on a warm gray background","model":"gpt-image-2","aspect_ratio":"1:1","hookUrl":"https://example.com/webhooks/openai-image"}'