Skip to content

Conversation

@rbren
Copy link
Contributor

@rbren rbren commented Oct 30, 2025

Description

This PR fixes an issue in the example application where the ConversationManager component was hardcoding gpt-4o-mini as the model name instead of using the user's configured settings.

Changes Made

  • Fixed hardcoded model: Replaced hardcoded 'gpt-4o-mini' with settings.modelName in the ConversationManager component
  • Removed base_url: Removed the hardcoded base_url: 'https://api.openai.com/v1' from the LLM configuration to allow users to use their preferred provider settings

Problem

The issue was introduced in commit 394ca3a47838c82afcae9e826b06d2bac7a81fa7 where the conversation creation logic was rewritten but accidentally hardcoded the model name instead of using the user's settings.

Impact

  • Users can now properly use their configured model name from the settings modal
  • Users can use different LLM providers without being forced to use OpenAI's base URL
  • The example app now correctly respects user preferences as intended

Testing

  • ✅ TypeScript compilation passes without errors
  • ✅ Vite build completes successfully
  • ✅ No breaking changes to the API or component interface

Files Changed

  • example/src/components/ConversationManager.tsx: Updated agent configuration to use user settings

This ensures the example application works as expected with user-configured settings rather than hardcoded values.

@rbren can click here to continue refining the PR

- Replace hardcoded 'gpt-4o-mini' with settings.modelName in ConversationManager
- Remove base_url from LLM configuration to use default provider settings
- Fixes issue introduced in commit 394ca3a

Co-authored-by: openhands <openhands@all-hands.dev>
@rbren rbren marked this pull request as ready for review October 30, 2025 15:27
@rbren rbren merged commit 95ee3a1 into main Oct 30, 2025
6 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants