Web user interaface does not parse model responses correctly

[kaanaksit]

LocalAI version:
Docker installation localai/localai: latest-gpu-nvidia-cuda-13

Environment, CPU architecture, OS, and Version:

$ uname -a
Linux COMPUTER 6.17.0-8-generic #8-Ubuntu SMP PREEMPT_DYNAMIC Fri Nov 14 21:44:46 UTC 2025 x86_64 GNU/Linux

Describe the bug
I installed LOCAL.AI using the docker option by following docker run -ti --name local-ai -p 8080:8080 --gpus all localai/localai:latest-gpu-nvidia-cuda-13. As I use the web user interface through localhost:8080, I encountered that the model responses that are provided in markdown is not parsed correctly.

To Reproduce

Expected behavior
Correctly parsed markdown.

Logs

sudo docker logs local-ai --tail 20
Jan 27 21:16:30 INFO  HTTP request method="GET" path="/static/components.css" status=200 
Jan 27 21:16:30 INFO  HTTP request method="GET" path="/static/animations.css" status=200 
Jan 27 21:16:30 INFO  HTTP request method="GET" path="/static/general.css" status=200 
Jan 27 21:16:30 INFO  HTTP request method="GET" path="/static/assets/fontawesome/css/brands.css" status=200 
Jan 27 21:16:30 INFO  HTTP request method="GET" path="/static/assets/space-grotesk-regular.ttf" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/assets/fontawesome/css/fontawesome.css" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/assets/playfair-display-bold.ttf" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/assets/tw-elements.css" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/assets/fontawesome/css/solid.css" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/assets/flowbite.min.js" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/assets/highlightjs.js" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/assets/alpine.js" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/assets/marked.js" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/assets/purify.js" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/chat.js" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/assets/tailwindcss.js" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/assets/pdf.min.js" status=200 
Jan 27 21:16:31 INFO  HTTP request method="GET" path="/static/logo_horizontal.png" status=200 
Jan 27 21:16:46 WARN  guessDefaultsFromFile: full offload is recommended
Jan 27 21:16:47 INFO  HTTP request method="POST" path="/v1/chat/completions" status=200

[kaanaksit]

The issue stands still in the passing time frame. But, as an intermediate solution, one can install Open Web UI using the following command in their terminal:

docker run -d -p 3000:8080 -v open-webui:/app/backend/data --name open-webui ghcr.io/open-webui/open-webui:main

and follow the following guideline to run it:

Connecting Open WebUI to LocalAI (Docker)

To bridge Open WebUI (port 3000) and LocalAI (port 8080) when both are in separate Docker containers, use a network address that allows the containers to communicate via the host.

1. Configure the Connection

1. Open Open WebUI at http://localhost:3000.
2. Go to Settings > Connections.
3. In the OpenAI API section, click the + icon.
4. Set the URL based on your OS:
   • Windows / macOS: http://host.docker.internal
   • Linux: http://localhost:8080 (Default Docker bridge)
5. API Key: Enter localai (placeholder or leave it as none).
6. Click Save.

2. Verify Models

• Refresh the connection.
• Your LocalAI models will now appear in the Model Selection dropdown at the top of the chat.

3. Troubleshooting

• Linux: Ensure the container was started with --add-host=host.docker.internal:host-gateway.
• Logs: Check the LocalAI Documentation if models fail to load.

Screenshot

[jaykayenn]

Any ideas? The output has been unusable since the latest update. All CRLF is missing.

Correction: It does not seem to parse Markdown at all.

[kaanaksit]

I don't have any relation to the LOCAL.AI project nor Open Web UI. But I would strongly recommend switching to Open Web UI for interface.