How to use Z-Image Turbo online

Oct 16, 2026

You can create an image in the browser without installing a local model. This guide describes the current image generator, including the signed-in free text-to-image option and the steps to find your results later.

1. Sign in and choose a mode

Open the generator and sign in. To use the current free option, choose Text to Image and Z-Image Turbo. This combination uses no credits for signed-in users. Other models and image-to-image use credits; see the free and credit usage guide before choosing.

For Image to Image, select an available model and upload your reference image. Use an image you have permission to upload. Reference files need to be added again if you leave for sign-in and return; the temporary input draft does not preserve those files.

2. Describe the image you want

Start with the subject, setting, lighting, style and composition. The three examples below use the prompts shown in homepage Image inspiration, with small changes you can try for a specific purpose.

Each link fills only the prompt in the generator's current language. It does not load the pictured image, select a model, change the mode or aspect ratio, or start generation. A restored login draft can take priority, so check the text after opening the link. Choose your model and ratio, review the cost, and explicitly choose Generate Images when ready.

Portrait: a natural courtyard headshot

Load the courtyard portrait prompt. The starting scene describes a fictional adult man in a light-blue linen shirt, framed from the chest up with his entire head visible, soft morning light and a blurred green courtyard.

  • For a profile image: keep the subject, shirt and light, choose a square ratio, and replace “Landscape 4:3” in the prompt with “Square composition, entire head and shoulders visible, space above the hair.” Keep the text and ratio control consistent.
  • If the background competes with the face: change only the background instruction to “A simple, softly blurred green background with no distinct objects.” Compare the framing and background before changing the lighting too.

This prompt describes a fictional person, not a way to reproduce a real person's identity. For edits to a photo you have permission to use, follow the reference-image guide.

Interior: a sunlit reading corner

Load the sunlit reading room prompt. Its main anchors are light-oak shelves, cream walls, one rounded armchair, a small table and afternoon light from the left window.

  • For a calmer room: retain those materials and the window position, but ask for “One armchair, one small table, mostly clear floor space and no extra decorative furniture.” This gives you a concrete object count to check.
  • If the room looks distorted: keep the furnishings unchanged and try “Eye-level camera, straight vertical lines, natural furniture proportions, no extreme wide-angle view.” Inspect the window frame and shelf edges in the next result.

Choose a landscape ratio in the control as well as describing it in the prompt. These are visual-design suggestions, not a measured floor plan or proof that furniture will fit a real room.

Still life: an amber fragrance bottle

Load the amber fragrance prompt. Start with one unbranded amber glass bottle, a matte cap, a cream stone pedestal, a small sprig of leaves and soft light passing through the glass.

  • For a cleaner product-style image: keep the bottle and pedestal, remove the leaves from the description, and request “One bottle only, full cap and base visible, ample space around the product.”
  • If bright reflections hide the bottle: change the lighting to “Soft diffused side light, restrained highlights, a clearly defined glass outline.” Check the outline, cap and base before adding more props.

Keep the no-label and no-logo instructions when you do not need branding. If exact packaging or readable label text matters, inspect it separately; the prompt does not guarantee product or text accuracy.

For all three scenes, change one element at a time and keep a copy of the previous prompt. These are suggested adjustments, not tested before-and-after results; the same prompt does not guarantee the pictured image.

3. Check the cost, then generate

Signed-in Z-Image Turbo text-to-image uses 0 credits. Paid image costs depend on the selected model and settings: check the quote and confirm the displayed credit cost before submitting. Existing credit balances, expiration dates and subscription records are preserved. Free generation is limited to 10 tasks per day, one image per task, within the site-wide US$10 monthly free budget (Taipei calendar month). When exhausted, free generation resumes next month.

Choose Generate Images once. The interface distinguishes submission, an accepted task waiting for a result, and processing. Processing time varies; the status does not guarantee a completion time.

If you are asked to sign in after entering text, the supported flow can temporarily restore your prompt and options in the same tab. Review the restored input before generating. Browser storage may be unavailable or the draft may expire, so keep your own copy of text you need to retain.

4. Review and download the result

When an image appears, use Download image below it. Downloading saves the returned image; it does not submit a new generation. Keep a local copy you need, since a provider image link may later expire.

To return later, use the account menu's Image tasks entry or open your image tasks. Sign in with the account that created the task. Open its result to view available images, download them, or return to the generator. Task history is private to the owning account.

5. Check the task before retrying

If the page cannot confirm the final status, check the task record first. A client timeout does not by itself prove the provider failed. Use the task's refresh action when offered, then review the resulting status.

A failed task, a task with no returned image and an unavailable image link are different situations. If a download fails, retry the download before submitting another generation. A new generation creates another task and can use credits under the selected mode's rules. This guide does not promise automatic refunds or permanent image storage.

Choose your next step

The homepage inspiration images show visual ideas, with prompts you can read and adapt. Review the selected model and settings before generating your own version.

For detailed checks, see reference uploads and login drafts and task status and download troubleshooting.