feat(llms): added additional params to llm-based blocks for alternative models#1223
feat(llms): added additional params to llm-based blocks for alternative models#1223waleedlatif1 merged 2 commits intostagingfrom
Conversation
![greptile-apps[bot]](https://avatars.githubusercontent.com/in/867647?s=60&v=4){kind=small}
There was a problem hiding this comment.
Greptile Summary
This PR enhances LLM-based blocks (router, evaluator, translate) to achieve feature parity with the agent block by adding comprehensive Azure OpenAI support and expanding model provider options. The changes introduce Azure-specific configuration parameters (azureEndpoint and azureApiVersion) that were previously missing, enabling proper Azure OpenAI integration. All affected blocks now support a broader range of model providers including Ollama, OpenRouter, and newer providers like XAI and DeepSeek.
Key architectural improvements include:
- Model Selection Enhancement: Upgraded from dropdown to combobox components with icon support for better UX
- Provider Expansion: Added support for OpenRouter models and newer chat providers (xai_chat, deepseek_chat, deepseek_reasoner)
- Conditional UI Logic: Implemented smart API key visibility that hides unnecessary fields for hosted environments and Ollama models
- Temperature Control: Added temperature parameters to router and evaluator blocks for consistent model behavior
- Tool Configuration Updates: Switched from
getBaseModelProviders()togetAllModelProviders()for comprehensive model support
The changes maintain backward compatibility while standardizing the configuration interface across all LLM blocks. This creates a unified experience where users can leverage the full spectrum of available AI models with proper provider-specific configuration options.
Confidence score: 3/5
- This PR introduces significant functionality but has potential runtime issues with provider state access patterns
- Score reflects complex conditional logic and function call patterns that could fail in edge cases
- Pay close attention to router.ts and translate.ts files which have the most complex provider state management logic
4 files reviewed, no comments
…ve models (#1223) * feat(llms): added additional params to llm-based blocks for alternative models * add hidden temp param to other LLM-based blocks
…ve models (simstudioai#1223) * feat(llms): added additional params to llm-based blocks for alternative models * add hidden temp param to other LLM-based blocks
1 participant
Summary
added additional params to llm-based blocks for alternative models, azure models were previously available but not all the params were exposed. Now there's full parity between the other LLM-based blocks and the agent block.
Type of Change
Testing
Tested manually.
Checklist