issue: Content errors when previewing AI-generated HTML and downloading via openWebUI

GiteaMirror commented

2025-11-11 16:53:22 -06:00

Owner

Originally created by @axh-xecoy on GitHub (Sep 12, 2025).

Check Existing Issues

I have searched for any existing and/or related issues.
I have searched for any existing and/or related discussions.
I am using the latest version of Open WebUI.

Installation Method

Docker

Open WebUI Version

v0.6.28

Ollama Version (if applicable)

v0.11.10

Operating System

Debian 12

Browser (if applicable)

Microsoft Edge 140.0.3485.54

Confirmation

I have read and followed all instructions in README.md.
I am using the latest version of both Open WebUI and Ollama.
I have included the browser console logs.
I have included the Docker container logs.
I have provided every relevant configuration, setting, and environment variable used in my setup.
I have clearly listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc).
I have documented step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation. My steps:
Start with the initial platform/version/OS and dependencies used,
Specify exact install/launch/configure commands,
List URLs visited, user input (incl. example values/emails/passwords if needed),
Describe all options and toggles enabled or changed,
Include any files or environmental changes,
Identify the expected and actual result at each stage,
Ensure any reasonably skilled user can follow and hit the same issue.

Expected Behavior

I am using the Qwen3 30B A3B (MOE) model in openWebUI and trying to make it generate a cute clock HTML.
The model used here is: ollama pull qwen3:30b-a3b
The prompt used here is: 请给我用单 html 页面编写一个可爱的时钟，有分针秒针时针，需要对应实际的时间

The AI has output the corresponding code. However, when I try to preview it in the page or directly download the HTML via the download button, there are clearly a lot of formatting errors in the downloaded HTML. But when I try to directly copy the code and paste it into an empty HTML file, the content is normal.

Actual Behavior

Under normal circumstances, when I download the file or preview it within the openWebUI page, the content and code should be consistent with the actual output from the AI, rather than having a large number of formatting errors.

Steps to Reproduce

Using the model obtained via "ollama pull qwen3:30b-a3b", I asked the model the question: "请给我用单 html 页面编写一个可爱的时钟，有分针秒针时针，需要对应实际的时间". So far, I have tried asking this question 3 times, and the same issue has occurred each time.

Logs & Screenshots

copy to file.html

artifact-939a3589-d705-4db2-a891-9cc284b8b9f8-0.html

docker-compose.yml

Additional Information

No response

Originally created by @axh-xecoy on GitHub (Sep 12, 2025). ### Check Existing Issues - [x] I have searched for any existing and/or related issues. - [x] I have searched for any existing and/or related discussions. - [x] I am using the latest version of Open WebUI. ### Installation Method Docker ### Open WebUI Version v0.6.28 ### Ollama Version (if applicable) v0.11.10 ### Operating System Debian 12 ### Browser (if applicable) Microsoft Edge 140.0.3485.54 ### Confirmation - [x] I have read and followed all instructions in `README.md`. - [x] I am using the latest version of **both** Open WebUI and Ollama. - [x] I have included the browser console logs. - [x] I have included the Docker container logs. - [x] I have **provided every relevant configuration, setting, and environment variable used in my setup.** - [x] I have clearly **listed every relevant configuration, custom setting, environment variable, and command-line option that influences my setup** (such as Docker Compose overrides, .env values, browser settings, authentication configurations, etc). - [x] I have documented **step-by-step reproduction instructions that are precise, sequential, and leave nothing to interpretation**. My steps: - Start with the initial platform/version/OS and dependencies used, - Specify exact install/launch/configure commands, - List URLs visited, user input (incl. example values/emails/passwords if needed), - Describe all options and toggles enabled or changed, - Include any files or environmental changes, - Identify the expected and actual result at each stage, - Ensure any reasonably skilled user can follow and hit the same issue. ### Expected Behavior I am using the Qwen3 30B A3B (MOE) model in openWebUI and trying to make it generate a cute clock HTML. The model used here is: `ollama pull qwen3:30b-a3b` The prompt used here is: `请给我用单 html 页面编写一个可爱的时钟，有分针秒针时针，需要对应实际的时间` The AI has output the corresponding code. However, when I try to preview it in the page or directly download the HTML via the download button, there are clearly a lot of formatting errors in the downloaded HTML. But when I try to directly copy the code and paste it into an empty HTML file, the content is normal. ### Actual Behavior Under normal circumstances, when I download the file or preview it within the openWebUI page, the content and code should be consistent with the actual output from the AI, rather than having a large number of formatting errors. ### Steps to Reproduce Using the model obtained via "ollama pull qwen3:30b-a3b", I asked the model the question: "请给我用单 html 页面编写一个可爱的时钟，有分针秒针时针，需要对应实际的时间". So far, I have tried asking this question 3 times, and the same issue has occurred each time. ### Logs & Screenshots <img width="2560" height="1540" alt="Image" src="https://github.com/user-attachments/assets/37bd2f38-4d8e-450e-ba66-f44c8589f708" /> <img width="2560" height="1540" alt="Image" src="https://github.com/user-attachments/assets/1d8c799e-d90d-40e4-a5c9-b382ff8e823c" /> <img width="2560" height="1540" alt="Image" src="https://github.com/user-attachments/assets/50056a0a-fa07-4317-8812-30f4a516bb5b" /> [copy to file.html](https://github.com/user-attachments/files/22291990/copy.to.file.html) [artifact-939a3589-d705-4db2-a891-9cc284b8b9f8-0.html](https://github.com/user-attachments/files/22291992/artifact-939a3589-d705-4db2-a891-9cc284b8b9f8-0.html) [docker-compose.yml](https://github.com/user-attachments/files/22292011/docker-compose.yml) ### Additional Information _No response_