🪄 AI Text to Image

Needs a third-party API · keys stored locally

Generate marketing visuals from prompts using your own model API.

The workbench controls are currently in Chinese; an English interface follows in the next batch.

Workbench coming soon

This tool is still in development. It is an AI tool, so once live it will connect to a third-party API you configure yourself. You can try the similar tools below in the meantime, or tell us which feature you need most via 「Feedback」 in the top bar.

How to use AI Text to Image

How it will work once live — see similar tools below in the meantime

1

Set up your model API

Open the page and start — no registration or sign-in needed.

2

Enter the image description and aspect ratio

Settings can be changed at any time; requests go straight from your browser to the API you configured.

3

Generate and download the image

Copy or download the result when you are done. Nothing is kept on this site.

Page Actions

Enter a description

Describe the image you want in Chinese or English (for example, "a stainless steel tumbler on a white background").

Choose style and size

Styles include product photography, illustration, and flat design, with aspect ratios such as 1:1 and 16:9.

Generate

Call the text-to-image model. Images appear in tens of seconds, and you can draw again several times.

Download / History

Save the images you like. Past results can be viewed again in the history.

How It Works & Highlights

How it works, and where it fits

This tool turns text descriptions into images through a third-party text-to-image API (you provide your own key). The key is stored in the local browser, and prompts do not pass through this site's server.

Highlights

  • ✓ Chinese prompts work directly, with no translation needed
  • ✓ Preset styles cover product images and posters
  • ✓ History keeps your ideas from getting lost

Tips

  • 💡 The more specific the prompt, the better: include material, colour, background, and lighting
  • 💡 For product images, "studio lighting, white background" gives the most reliable results
  • 💡 Generate several images, pick the best one, and refine it, instead of struggling over a single attempt

FAQ

Privacy, format limits, and usage tips

It depends on the third-party text-to-image API you configure. The tool has no built-in model. The API address and key are provided by you and stored only in the local browser.
It depends on the terms of service of the model you use. Confirm the licensing terms of the provider before commercial use. This site makes no guarantee.
Yes. Requests are relayed by the TWC AI gateway to the provider you configured, and the content is not retained. With your own key (BYOK), the key is configured on the account page and stored encrypted in this site's database (D1), not in browser local storage; deleting the configuration removes it.
Any recent browser works (Chrome, Edge, Firefox, or Safari from the last three years). It works on phones too, but batch file handling and table checks are more efficient on a large screen.
Refresh the page first (resources may be cached by the browser), then check that the file format and size match the hints on the page. If it still fails, send the steps to reproduce via 「Feedback」 in the top bar.
Limited by the third-party API you configured — the provider's quota, concurrency, and per-request limits apply.
AI tools are limited by API quota and concurrency. Process in small batches to avoid rate limits or timeouts.
Results are mainly text or images shown on the page, which can be copied or downloaded. Follow the workbench buttons.
No. It needs to reach the third-party API you configured, so it cannot generate results offline.
No. Favorites and recent-use records only store tool names, never your files or text. When you are not signed in, favorites are kept only in your browser's localStorage; after signing in they sync to your account and work across devices. Recent-use records stay on your own machine.
First check that the input is complete and the format matches the hints on the page — you can click the example on the page to verify quickly. If it is still wrong, refresh the page and send us the steps and a screenshot via the feedback form.
Send your question or suggestion via 「Feedback」 in the top bar. Describe the use case and the outcome you want, and we will assess it before updating the tool and the page.