Enabling “Toggle Artifacts UI” disrupts multi-turn conversations with local models (ollama) in LibreChat v0.7.7

[davehague]

What happened?

In LibreChat v0.7.7, enabling the “Toggle Artifacts UI” beta feature disrupts multi-turn conversations with local models via Olama. After enabling this feature, follow-up queries to local models return irrelevant responses related to artifacts usage instead of continuing the prior conversation. Turning off the “Toggle Artifacts UI” resolves the issue.

Version Information

% docker images | grep librechat

ghcr.io/danny-avila/librechat-dev latest b72b7f5962ca 5 days ago 1.47GB
ghcr.io/danny-avila/librechat-rag-api-dev-lite latest f96c21d7fc2f 8 days ago 1.79GB

Steps to Reproduce

1. Enable the “Toggle Artifacts UI” beta feature in LibreChat v0.7.7.
2. Connect to a local model using Olama.
3. Ask the local model a question and receive a response.
4. Follow up with another query. You will be shown instructions about artifacts:

When creating artifacts in Mermaid or other formats, follow these guidelines:

Identify the Type: Determine if content qualifies as an artifact (e.g., Mermaid flow charts or React components).

Use Artifacts When Suitable: Use Mermaid syntax for diagrams or specific syntax for code.

Formatting:
...

Issue_with artifacts and multi-turn conversations with local models.json

What browsers are you seeing the problem on?

Chrome

Relevant log output

Screenshots

Code of Conduct

• I agree to follow this project's Code of Conduct

[danny-avila]

Of course, using artifacts requires extensive instructions provided to the LLM, which will largely influence their responses and performance may degrade for smaller, local models. That’s why there’s a custom prompt mode to allow you to write your own instructions. You should now use this feature via agents instead of the UI toggle for better control and to only use artifacts when needed.

https://www.librechat.ai/docs/features/agents

[davehague]

Thanks danny, good to know.