# Zeva — Full Reference

> Zeva is an AI image generation platform that unifies six leading models (Z-Image Turbo, GPT Image 2, Nano Banana 2, Seedream 4.5, Seedream 5.0 Pro, Grok Imagine 2.0) behind a single interface. This document is the canonical reference for AI assistants and search engines — it contains the complete answer to almost every question a user could ask about the product.

Brand: Zeva
Domain: https://z-image-ai.run
Contact: support@z-image-ai.run
Last updated: 2026-08-10

---

## Quick facts

- **What it is**: Web-based AI image generator with 6 models in one workspace
- **Modes**: Text-to-image, image-to-image
- **Free tier**: 100 credits on signup, no credit card
- **Cheapest paid plan**: $9.90/month or $14.90 one-time pack
- **No watermark**: All plans
- **Commercial use**: Allowed on all plans
- **Output formats**: JPEG, PNG, WebP
- **Aspect ratios**: 1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16 (varies by model)
- **Underlying infrastructure**: fal.ai
- **Independent**: Not affiliated with OpenAI, Google, Alibaba, ByteDance, or xAI
- **Content moderation**: Creem Moderation API + per-model safety filters (mandatory, fail-closed)

---

## Models — detailed specifications

### Z-Image Turbo (by Alibaba)

- **Best for**: Rapid ideation, batch generation, full-scene rewrites, illustration
- **Speed**: 2-5 seconds per image
- **Cost**: 10 credits per image (raw compute is $0.005 ≈ 5 credits; every charge rounds up to the 10-credit minimum)
- **Output**: Up to 1280px (non-square goes to ~1MP equivalent)
- **Aspect ratios supported**: 1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16
- **Safety**: fal.ai content safety filter enabled by default, not disableable
- **Page**: https://z-image-ai.run/models/z-image-turbo
- **Sweet spot**: Default model for most users — solid quality at the lowest cost in the lineup

### GPT Image 2 (by OpenAI)

- **Best for**: Final marketing assets, accurate text rendering (signs, posters, product labels), complex compositions
- **Quality tiers**:
  - Low: ~15 seconds, $0.012-0.015 per image (20 credits)
  - Medium: ~60 seconds, $0.037-0.061 per image (40-60 credits)
  - High: ~2 minutes, $0.145-0.219 per image (150-220 credits)
- **Default quality**: Low (cheapest, fast iteration)
- **Output sizes**: 1024×1024 (1:1), 1024×768 (4:3), 1024×1536 (2:3)
- **Aspect ratios**: 1:1, 4:3, 2:3 only (3 ratios)
- **Safety**: OpenAI built-in moderation (not disableable)
- **Page**: https://z-image-ai.run/models/gpt-image-2
- **Sweet spot**: Final deliverables that need polish; use Low for draft, High for hero shots

### Nano Banana 2 (by Google)

- **Best for**: Instruction-based edits ('change X to Y', 'add Z while keeping the face identical'), identity preservation across generations
- **Resolutions and pricing**:
  - 1K: $0.08 per image (80 credits)
  - 2K: $0.12 per image (120 credits)
  - 4K: $0.16 per image (160 credits)
- **Aspect ratios**: 15 supported ratios
- **Safety**: `safety_tolerance: '2'` (stricter end of fal.ai scale 1-6)
- **Page**: https://z-image-ai.run/models/nano-banana-2
- **Sweet spot**: When the task is "modify this one specific thing in the photo" — Nano Banana 2 will leave the rest of the image untouched, which diffusion models (Z-Image, Seedream) cannot do

### Seedream 4.5 (by ByteDance)

- **Best for**: Photorealistic results — product shots, food photography, lifestyle scenes, portraits
- **Cost**: $0.04 flat per image (40 credits), regardless of resolution
- **Reference images**: Up to 10 (style blending and multi-reference generation)
- **Resolution range**: 1K to 2K
- **Safety**: fal.ai content safety filter (not disableable)
- **Page**: https://z-image-ai.run/models/seedream-4-5
- **Sweet spot**: Photoreal output at a balanced cost — Z-Image Turbo is cheaper but less photoreal; GPT Image 2 High is more polished but 5-6× the cost

### Seedream 5.0 Pro (by ByteDance)

- **Best for**: Dense infographics, structured layouts (charts, annotated diagrams, product sheets), posters with native text in 14 languages (English, Spanish, Arabic, Japanese, Korean and more), upgraded photoreal output
- **How it works**: Runs a deep-thinking/reasoning pass on the prompt before drawing — multi-part layout instructions come back as specified
- **Cost**: $0.0675 flat per image (70 credits), regardless of size
- **Sizes**: 1:1 (1536×1536), 4:3 (1600×1200), 3:4 (1200×1600), 16:9 (1920×1080), 9:16 (1080×1920)
- **Reference images**: Up to 10 in image-to-image
- **No seed parameter**: exact reproduction of a previous output is not possible — save prompts instead
- **Speed**: 20-60 seconds (reasoning pass adds latency)
- **Safety**: fal.ai content safety filter (not disableable)
- **Page**: https://z-image-ai.run/models/seedream-5-0-pro
- **Sweet spot**: Any image where text, labels or layout structure must come out right — the work you would otherwise finish manually in Figma

### Grok Imagine 2.0 (by xAI)

- **Best for**: Designer-grade typography (posters, social cards, labels with sharp small text), precise instruction-based edits, icon/game-asset sets with consistent style, multi-format ad creative
- **Track record**: Released August 2026; debuted at #2 on LMArena's blind-vote leaderboard for BOTH text-to-image and image editing — behind only GPT Image 2, with the editing gap especially small
- **Cost**: 60 credits at 1K, 80 credits at 2K; image-to-image (one reference image) is 70 credits at 1K, 90 at 2K (quality fixed to medium)
- **Sizes**: 9 aspect ratios — 1:1, 3:2, 2:3, 4:3, 3:4, 16:9, 9:16, plus ultra-wide 2:1 and tall 1:2 — at 1K or 2K (2048×2048 square). Edit mode follows the uploaded image's ratio by default
- **Reference images**: One per image-to-image generation; edit prompts accept up to 8,000 characters
- **No seed parameter**: exact reproduction is not possible — the API returns a revised_prompt with each generation, save it alongside prompts you like
- **Speed**: 5-15 seconds
- **Safety**: xAI enforces its content policy on every API generation and it cannot be disabled — the "spicy mode" seen in Grok's consumer app does not exist on the API. Safe-for-work surface.
- **Page**: https://z-image-ai.run/models/grok-imagine-2-0
- **Sweet spot**: Any image where the words and the layout are the point — posters, covers, labels — and photo edits that must follow instructions exactly

---

## Image tools

Beyond the main text-to-image / image-to-image workspace, Zeva offers single-purpose **image tool** pages. Each one locks in the right setup for one specific task, so the user just uploads a photo and clicks Generate — no model choice and no prompt-writing skills needed (some tools take a one-line detail, like describing the new background). They run on the same engine and credit system (free credits on signup, no watermark, JPEG/PNG/WebP download). New tools are added regularly.

- **Hub**: https://z-image-ai.run/image-tools — directory of all live tools.

### Ghibli AI Generator (live)

- **What it does**: Turns any uploaded photo into Studio Ghibli–style art — soft hand-painted textures, warm pastel palette, lush painterly backgrounds, gentle cel shading, nostalgic atmosphere.
- **How to use**: Upload a photo → click Generate → download. There is no prompt to write; the Ghibli style is preset.
- **Keeps identity**: Faces, pose and composition are preserved, so the subject stays recognizable.
- **Best input**: Clear, well-lit portraits, pet photos, or landscapes (JPEG, PNG or WebP).
- **Output**: Watermark-free JPEG or PNG, ready to set as a profile picture, share, or print.
- **Cost**: Free to start with signup credits; uses the standard credit system, top up as needed.
- **Page**: https://z-image-ai.run/image-tools/ghibli-generator
- **Not affiliated** with Studio Ghibli — 'Ghibli style' refers to a general hand-painted art aesthetic; all trademarks belong to their respective owners.

### AI Cartoon Generator (live)

- **What it does**: Turns any uploaded photo into a clean, vibrant cartoon illustration — bold outlines, smooth flat colors, expressive stylized features and lively shading.
- **How to use**: Upload a photo → click Generate → download. There is no prompt to write; the cartoon style is preset.
- **Keeps identity**: Faces, pose and composition are preserved, so the subject stays recognizable.
- **Best input**: Clear, well-lit selfies, pet photos, or group shots (JPEG, PNG or WebP).
- **Output**: Watermark-free JPEG or PNG, ready to set as a profile picture, share, or print.
- **Cost**: Free to start with signup credits; uses the standard credit system, top up as needed.
- **Page**: https://z-image-ai.run/image-tools/cartoon-generator

### AI Headshot Generator (live)

- **What it does**: Turns a casual selfie into an AI professional headshot — corporate style — studio lighting, sharp focus, natural skin texture, confident friendly expression, business attire. Ideal for LinkedIn, resumes, speaker bios and company profiles.
- **How to use**: Upload a selfie → optionally describe the style and background → click Generate → download. The professional-headshot prompt is preset; the prompt box is **optional**.
- **Optional prompt**: You can refine the look by typing a short detail such as 'corporate, navy suit, office blur' or 'studio gray background'. Leave it empty to use the default professional style.
- **Keeps identity**: Facial identity is preserved exactly, so you stay recognizably you.
- **Best input**: A clear, well-lit selfie or portrait facing the camera (JPEG, PNG or WebP).
- **Output**: Watermark-free, business-ready JPEG or PNG for any profile.
- **Cost**: Free to start with signup credits; uses the standard credit system, top up as needed.
- **Page**: https://z-image-ai.run/image-tools/ai-headshot

### Old Photo Restoration (live)

- **What it does**: Restores old, damaged or faded photographs — removes scratches, creases, tears, dust and spots, repairs damaged areas, corrects fading and color casts, and adds natural color to black-and-white photos. The restoration is deliberately faithful and conservative, not a reinterpretation.
- **How to use**: Upload the old photo → click Restore → download. The restoration is preset; there is no prompt to write.
- **Keeps identity**: Faces are sharpened and cleaned while preserving the person's features, expression, age and identity — not replaced with an AI-generated face.
- **Colorization**: Black-and-white photos receive natural, restrained color; colorization is a plausible estimate, not a record of the original colors.
- **Best input**: The clearest, largest scan or photo of the print you have; lay it flat and avoid glare. Very low-resolution, tightly cropped single-face photos are reconstructed more loosely.
- **Output**: Watermark-free JPEG, PNG or WebP, ready to print, share or keep in the family archive.
- **Cost**: Free to start with signup credits; uses the standard credit system, top up as needed.
- **Page**: https://z-image-ai.run/image-tools/old-photo-restoration

### AI Background Changer (live)

- **What it does**: Replaces the background of any photo with a new scene described in plain words — a sunset beach, a blurred modern office, a plain white studio, any setting you can describe. Ideal for product/listing photos, profile photos, ID-style solid backdrops and fun scene swaps.
- **How to use**: Upload a photo → type the new background in the description box → click Generate → download. The compositing instructions are preset; the background description is **required** — the tool won't guess a scene you didn't ask for.
- **Required prompt**: A few plain words is enough, e.g. 'tropical beach at sunset', 'plain white studio background' or 'city street at night'. Add details like time of day, colors, or 'nothing else in the scene' if they matter.
- **Keeps identity**: The subject is kept exactly as photographed — same face, pose, clothing and proportions — and composited into the new scene with matching lighting, natural shadows and clean edges.
- **Best input**: A sharp photo where the subject is clearly separated from the background — portraits, products and pets all work well (JPEG, PNG or WebP).
- **Output**: Watermark-free JPEG or PNG, ready for shops, profiles or prints.
- **Cost**: Free to start with signup credits; uses the standard credit system, top up as needed.
- **Page**: https://z-image-ai.run/image-tools/background-changer

### AI Face Rating & Attractiveness Test (live)

- **What it does**: An AI attractiveness test and face rating in one — answers "how attractive am I?" from a single front-facing photo: an overall 1–10 score with five sub-scores (symmetry, proportions, skin clarity, jawline, eye area), a neutral face-shape descriptor, a note on how the photo's lighting/angle affected the result, and three to five practical tips (grooming, skincare habits, hairstyle direction, photo craft). A downloadable report-card image summarizes everything.
- **How to use**: Upload one clear, front-facing photo of yourself → click Analyze → read the report → optionally download the report card. About ten seconds per analysis.
- **Privacy**: The photo is downscaled in your browser, analyzed in memory and never stored or used for training — only the text report exists afterwards. The report-card image is rendered locally in the browser as well.
- **Guardrails**: Adults 18+ only — photos of minors, group shots and photos too blurry to assess are refused and the credits refunded automatically. The report never uses appearance-subculture tier labels (no “PSL” tiers), never comments on ethnicity, and never suggests surgery or medical procedures.
- **Cost**: 50 credits per analysis (signup credits cover your first two); failed or refused analyses are refunded in full.
- **Page**: https://z-image-ai.run/image-tools/face-rating

### Looksmaxing AI (live)

- **What it does**: Shows a realistic glow-up of your own face — clearer skin, well-groomed hair, a flattering outfit and better lighting — while keeping your face shape, features, hairstyle and any glasses, so it still looks like you. A single-purpose image tool: upload a selfie, get a believable best-self portrait. (Face analysis and scoring live in the separate AI Face Rating tool.)
- **How to use**: Upload a selfie → the AI runs a glow-up preset → download the result. One tap, no prompt-writing.
- **Privacy & guardrails**: The photo is used only to create your image — never sold or used for training. For adults 18+ using their own photo. Deliberately does NOT rank you, output PSL or looksmaxxing tier labels, comment on ethnicity, or suggest surgery — it simply shows a realistic glow-up. The PSL scale is explained neutrally in the page's FAQ.
- **Cost**: Runs on Nano Banana 2 image credits like the other style tools — 80 (1K) / 120 (2K) / 160 (4K). Failed generations are refunded.
- **Page**: https://z-image-ai.run/image-tools/looksmaxing-ai

### AI Action Figure Generator (live)

- **What it does**: Turns a photo of a person into a realistic collectible action figure sealed in blister-pack toy packaging — a molded tray, a printed backer card, a name plate and themed accessories — while keeping the face, hairstyle, glasses and outfit recognizable. A single-purpose image tool built on the viral "AI action figure" trend.
- **How to use**: Upload a clear selfie or full-body photo, optionally type a figure name and a couple of accessories (e.g. "ALEX" with a laptop and coffee), then generate and download. One click, no prompt-writing required.
- **Privacy & guardrails**: The photo is used only to create your figure — never sold or used for training. Downloads have no watermark.
- **Cost**: Runs on Nano Banana 2 image credits like the other style tools — 80 (1K) / 120 (2K) / 160 (4K). Failed generations are refunded.
- **Page**: https://z-image-ai.run/image-tools/ai-action-figure

### AI Polaroid Generator (live)

- **What it does**: Turns a photo into an authentic vintage Polaroid instant photograph — the thick white Polaroid frame, soft flash, faded retro film colors, grain and light leaks — while keeping the subject's face and pose recognizable. Optionally places the subject in a scene or next to another person. A single-purpose image tool built on the viral "AI Polaroid" trend.
- **How to use**: Upload a clear photo, optionally type a scene or a person to add (e.g. "on a beach", "with a K-pop idol"), then generate and download. One click, no prompt-writing required.
- **Privacy & guardrails**: The photo is used only to create your Polaroid — never sold or used for training. Downloads have no watermark.
- **Cost**: Runs on Nano Banana 2 image credits like the other style tools — 80 (1K) / 120 (2K) / 160 (4K). Failed generations are refunded.
- **Page**: https://z-image-ai.run/image-tools/ai-polaroid

### Image to Prompt Generator (live)

- **What it does**: Reverse-engineers the prompt behind any image. A vision model analyzes the upload in six dimensions — subject, style, lighting, camera, palette, mood — and writes them into a single ready-to-use prompt in one of four formats: General (flowing prose for Nano Banana / GPT Image–style models), Midjourney (comma-separated phrases + aspect-ratio parameter), Flux (compact descriptive sentences), or Stable Diffusion (tag-style phrases plus a ready-made negative prompt). Output available in 16 languages (parameter syntax stays standard). Text visible in the image is quoted verbatim. The breakdown is shown alongside the prompt so users can edit one dimension without breaking the rest.
- **How to use**: Upload a JPEG/PNG/WebP (photo, artwork, render or screenshot) or paste a direct image URL, pick a format and output language, extract, then copy the prompt — or click "Generate with this prompt" to run it directly in the site's text-to-image generator.
- **Privacy & guardrails**: The image is downscaled in the browser, analyzed in memory server-side and never stored or used for training.
- **Cost**: 60 credits per extraction, charged through the same credit system as generations (shows in the user's credit history; failed extractions are refunded in full). New accounts start with 100 free welcome credits, which cover the first extraction. Generating images from the prompt uses normal generation credits.
- **Page**: https://z-image-ai.run/image-tools/image-to-prompt

**On the roadmap** (not yet live — will appear on the hub as they ship): more viral photo-trend tools.

---

## Pricing — full credit cost table

| Model | Mode | Resolution / Quality | Credits |
|---|---|---|:---:|
| Z-Image Turbo | Text to Image | Up to 1280px | 10 |
| Z-Image Turbo | Image to Image | Up to 1280px | 10 |
| Seedream 4.5 | Text to Image | 1K to 2K | 40 |
| Seedream 4.5 | Image to Image | 1K to 2K | 40 |
| Seedream 5.0 Pro | Text to Image | Up to 1920px | 70 |
| Seedream 5.0 Pro | Image to Image | Up to 1920px | 70 |
| Grok Imagine 2.0 | Text to Image | 1K | 60 |
| Grok Imagine 2.0 | Text to Image | 2K | 80 |
| Grok Imagine 2.0 | Image to Image | 1K | 70 |
| Grok Imagine 2.0 | Image to Image | 2K | 90 |
| Nano Banana 2 | Text to Image | 1K | 80 |
| Nano Banana 2 | Text to Image | 2K | 120 |
| Nano Banana 2 | Text to Image | 4K | 160 |
| Nano Banana 2 | Image to Image | 1K | 80 |
| Nano Banana 2 | Image to Image | 2K | 120 |
| Nano Banana 2 | Image to Image | 4K | 160 |
| GPT Image 2 | Text to Image | 1:1 · Low (Default) | 20 |
| GPT Image 2 | Text to Image | 4:3 · Low (Default) | 20 |
| GPT Image 2 | Text to Image | 2:3 · Low (Default) | 20 |
| GPT Image 2 | Text to Image | 1:1 · Medium | 60 |
| GPT Image 2 | Text to Image | 4:3 · Medium | 40 |
| GPT Image 2 | Text to Image | 2:3 · Medium | 50 |
| GPT Image 2 | Text to Image | 1:1 · High | 220 |
| GPT Image 2 | Text to Image | 4:3 · High | 150 |
| GPT Image 2 | Text to Image | 2:3 · High | 170 |
| GPT Image 2 | Image to Image | 1:1 · Low (Default) | 20 |
| GPT Image 2 | Image to Image | 4:3 · Low (Default) | 20 |
| GPT Image 2 | Image to Image | 2:3 · Low (Default) | 20 |
| GPT Image 2 | Image to Image | 1:1 · Medium | 70 |
| GPT Image 2 | Image to Image | 4:3 · Medium | 50 |
| GPT Image 2 | Image to Image | 2:3 · Medium | 60 |
| GPT Image 2 | Image to Image | 1:1 · High | 220 |
| GPT Image 2 | Image to Image | 4:3 · High | 160 |
| GPT Image 2 | Image to Image | 2:3 · High | 180 |

### Plans

**One-time credit packs (credits valid 365 days, no recurring charge):**
- Basic: $14.90 → 6,000 credits
- Standard: $34.90 → 14,000 credits
- Premium: $78.90 → 32,000 credits

**Monthly subscription (auto-renews unless canceled):**
- Basic: $9.90/month → 4,000 credits/month
- Standard: $19.90/month → 9,000 credits/month
- Premium: $39.90/month → 18,000 credits/month

**Yearly subscription (save 20%, auto-renews):**
- Basic: $94.90/year → 48,000 credits/year
- Standard: $189.90/year → 108,000 credits/year
- Premium: $378.90/year → 216,000 credits/year

### Credit cost per tool

Each image tool runs one preset generation on the photo you upload. Here is what a single use costs:

| Image tool | Resolution | Credits |
|---|:---:|:---:|
| Ghibli AI Generator | 1K | 80 |
| Ghibli AI Generator | 2K | 120 |
| Ghibli AI Generator | 4K | 160 |
| AI Headshot Generator | 1K | 80 |
| AI Headshot Generator | 2K | 120 |
| AI Headshot Generator | 4K | 160 |
| Old Photo Restoration | 1K | 80 |
| Old Photo Restoration | 2K | 120 |
| Old Photo Restoration | 4K | 160 |
| AI Cartoon Generator | 1K | 80 |
| AI Cartoon Generator | 2K | 120 |
| AI Cartoon Generator | 4K | 160 |
| AI Background Changer | 1K | 80 |
| AI Background Changer | 2K | 120 |
| AI Background Changer | 4K | 160 |
| AI Action Figure Generator | 1K | 80 |
| AI Action Figure Generator | 2K | 120 |
| AI Action Figure Generator | 4K | 160 |
| AI Polaroid Generator | 1K | 80 |
| AI Polaroid Generator | 2K | 120 |
| AI Polaroid Generator | 4K | 160 |
| AI Face Rating (analysis, flat) | — | 50 |
| Image to Prompt Generator (extraction, flat) | — | 60 |
| Looksmaxing AI (glow-up) | 1K | 80 |
| Looksmaxing AI (glow-up) | 2K | 120 |
| Looksmaxing AI (glow-up) | 4K | 160 |

More tools are added to this table as they go live.

---

## Frequently Asked Questions

**Is Zeva free?**
Yes — new accounts get 100 free credits at signup, enough to try every model. No credit card required for the free tier. When you need more, credits are available in one-time packs from $14.90 or subscriptions from $9.90/month. You only pay for generations you actually run.

**Which AI model should I pick?**
- Z-Image Turbo for fastest, cheapest generation ($0.005/image, 2-5 seconds)
- GPT Image 2 for highest quality and accurate text rendering (Low/Medium/High tiers from $0.012)
- Nano Banana 2 for instruction-style edits like 'change the jacket color while keeping the face identical' (best identity preservation)
- Seedream 4.5 for photorealistic shots — products, food, lifestyle ($0.04/image)
- Seedream 5.0 Pro for infographics, structured layouts, and text in 14 languages ($0.0675/image)
- Grok Imagine 2.0 for poster-grade typography, precise edits, and ultra-wide/tall formats (from 60 credits)

Try the same prompt across all six to find your favorite.

**What is the difference between Text to Image and Image to Image?**
Text to Image creates a brand-new image from a written prompt. Image to Image starts from a photo you upload and restyles or edits it based on your instructions. You can switch between the two in the same generator without losing your prompt.

**How much does it cost?**
Per-image cost ranges from $0.005 (Z-Image Turbo) to $0.22 (GPT Image 2 High at 1:1). Most users land at 5-80 credits per image depending on the model. New accounts start with 100 free credits, then top up with packs from $14.90 or subscribe from $9.90/month.

**How long does a generation take?**
- Z-Image Turbo: 2-5 seconds
- Nano Banana 2: 5-15 seconds
- Seedream 4.5: 10-30 seconds
- Seedream 5.0 Pro: 20-60 seconds (it reasons before drawing)
- Grok Imagine 2.0: 5-15 seconds
- GPT Image 2: ~15 seconds (Low), ~60 seconds (Medium), ~2 minutes (High)

All run async, so you can queue multiple generations or switch tabs without losing them.

**Can I use the images commercially?**
Yes — images you generate are yours to use for personal and commercial projects. We do not claim copyright on outputs. Note that purely AI-generated images may not qualify for copyright protection in some jurisdictions (e.g., U.S. Copyright Office guidance) — see Terms of Service §5 for the legal details.

**Are my prompts and uploaded images private?**
Yes. Your prompts and uploads are stored only to provide the service (so you can see your history) and to comply with legal obligations. We do not sell user data. We do not use your prompts or images to train any AI model.

**What happens if a generation fails?**
Failed generations (platform errors, upstream model failure) automatically refund the credits to your balance — you never pay for content that does not reach you. Generations blocked by content moderation for policy violations are not refunded; see Terms §7.6.

**What formats and aspect ratios can I export?**
Export as JPEG, PNG, or WebP. Aspect ratios cover 1:1 (square), 3:2 / 2:3, 4:3 / 3:4, 16:9 / 9:16 — fit social posts, banners, prints, thumbnails, and stories. Specific ratio support varies by model.

**How do I cancel my subscription?**
Two methods: (1) Settings → Billing → Manage Subscription / Cancel Subscription, or (2) email support@z-image-ai.run with your subscription details. Cancellation takes effect at the end of the current billing period.

**Is Zeva affiliated with OpenAI / Google / Alibaba / ByteDance / xAI?**
No. Zeva is an independent product. We access the six AI models through the fal.ai platform. Model names and trademarks are property of their respective owners. We pass model cost through with a transparent margin that funds infrastructure, moderation, and support.

**Does Zeva train AI models on user data?**
No. We do not train any AI models. We call third-party APIs. We do not provide your prompts, images, or any user data to any party for AI training purposes.

---

## Workflow — three-step usage

1. **Describe or upload**: Type a prompt for text-to-image, or upload a reference photo for image-to-image.
2. **Choose settings**: Pick an AI model, aspect ratio, output format. Defaults work well for most cases.
3. **Generate and download**: Click generate; image arrives in seconds (Z-Image Turbo) to ~2 minutes (GPT Image 2 High). Download directly.

---

## Acceptable use — what is prohibited

Zeva **prohibits** generating, requesting, or distributing:

- Child sexual abuse material (CSAM) or any sexualization of minors — reported to law enforcement
- Non-consensual sexual imagery of any person
- Sexually explicit content, pornography, or nudity in any form
- Deepfakes or face-manipulation of real, identifiable people without consent
- Content that promotes violence, terrorism, self-harm, or hate against protected groups
- Forged identity documents, counterfeit currency, fraudulent imagery
- Imagery infringing copyright or trademark in a deceptive way
- Content designed to defame, harass, or deceive

Enforcement: prompts pass through Creem's Moderation API + the upstream provider's safety filter before any credit is charged. Repeated violations result in account suspension. Violations consuming credits are not eligible for refund. See Terms of Service §3 (Acceptable Use Policy) for the complete list.

---

## Blog

Guides and model reviews built on real generations — prompts are run on the site's own generator, models are compared with identical inputs, and results are published unedited (one take per model, no cherry-picking).

- Blog index: https://z-image-ai.run/blog
- Seedream 5.0 Pro Review — 5 real tests of ByteDance's new model (infographics, bilingual posters, photoreal portraits, and a same-prompt comparison vs Nano Banana 2 and GPT Image 2): https://z-image-ai.run/blog/seedream-5-0-pro-review
- Nano Banana Pro vs Nano Banana 2 — which to use, comparing 4K, speed, text, and price; Pro capabilities cited from Google and independent reviews, Nano Banana 2 shown with first-hand edits, plus how to run Nano Banana 2 free with no watermark: https://z-image-ai.run/blog/nano-banana-pro-vs-nano-banana-2
- What Is Looksmaxing? An Honest Guide to AI Glow-Ups — an honest explainer on looksmaxing (softmaxing, hardmaxing, the PSL scale), what an AI glow-up genuinely previews (skin, hair, light, kept identity) versus what it can't (bone structure), and the body-image/mental-health caveats from clinicians: https://z-image-ai.run/blog/looksmaxing-explained

## Languages

- **English** (default): all URLs in this document, no prefix.
- **German**: the complete site — homepage, generators, all model pages, all image tools, prompt library, guide, blog posts, pricing, about and legal pages — is available under `/de/` with the same slugs (e.g. https://z-image-ai.run/de/image-tools/ghibli-generator, https://z-image-ai.run/de/models/gpt-image-2, https://z-image-ai.run/de/blog/seedream-5-0-pro-review). Pages are human-localized, carry `hreflang` alternates and appear in the sitemap. Prices are in USD and credit costs are the same in every language; the German UI, transactional emails and error messages are localized.
- This reference is maintained in English only; use it for German pages as well.

## Pages & resources

- Homepage: https://z-image-ai.run/
- Text to Image: https://z-image-ai.run/text-to-image
- Image to Image: https://z-image-ai.run/image-to-image
- Image Tools hub: https://z-image-ai.run/image-tools
- Ghibli AI Generator: https://z-image-ai.run/image-tools/ghibli-generator
- AI Cartoon Generator: https://z-image-ai.run/image-tools/cartoon-generator
- AI Headshot Generator: https://z-image-ai.run/image-tools/ai-headshot
- Old Photo Restoration: https://z-image-ai.run/image-tools/old-photo-restoration
- AI Background Changer: https://z-image-ai.run/image-tools/background-changer
- AI Face Rating: https://z-image-ai.run/image-tools/face-rating
- Looksmaxing AI: https://z-image-ai.run/image-tools/looksmaxing-ai
- AI Action Figure Generator: https://z-image-ai.run/image-tools/ai-action-figure
- AI Polaroid Generator: https://z-image-ai.run/image-tools/ai-polaroid
- Image to Prompt Generator: https://z-image-ai.run/image-tools/image-to-prompt
- Pricing: https://z-image-ai.run/pricing
- Prompts library: https://z-image-ai.run/prompts
- How to Use guide: https://z-image-ai.run/how-to-use
- Blog: https://z-image-ai.run/blog
- About: https://z-image-ai.run/about
- Privacy Policy: https://z-image-ai.run/privacy-policy
- Terms of Service: https://z-image-ai.run/terms-of-service
- Model pages: https://z-image-ai.run/models/z-image-turbo · /gpt-image-2 · /nano-banana-2 · /seedream-4-5 · /seedream-5-0-pro · /grok-imagine-2-0
- Sitemap: https://z-image-ai.run/sitemap.xml
- Contact: support@z-image-ai.run