We have replaced the internal LLM provider dependency cnoe-agent-utils with litellm to enable support for a wider range of LLM providers.
Impact
- This is a breaking change.
- Environment variables for LLM credentials must be updated in your environment to match the litellm convention for your preferred LLM provider. For a comprehensive list of supported providers, see the official litellm documentation.
- LLM provider and model must be specified by the LLM_MODEL env variable. Examples:
LLM_MODEL="openai/<model_of_choice>"
OPENAI_API_KEY=<your_openai_api_key>LLM_MODEL="azure/<your_deployment_name>"
AZURE_API_BASE=https://your-azure-resource.openai.azure.com/
AZURE_API_KEY=<your_azure_api_key>
AZURE_API_VERSION=<your_azure_api_version>LLM_MODEL="groq/<model_of_choice>"
GROQ_API_KEY=<your_groq_api_key>- Corto Demo: Completed the implementation of the Corto demo, including:
- Frontend:
- Developed a React-based frontend for the Corto Exchange
- Backend:
- Implemented the Corto Exchange backend with modules for:
- SLIM transport integration
- A2A client
- Graph operations
- Main entry point
- Developed the Corto Farm backend with:
- SLIM transport integration
- A2A agent execution and card management
- Farm server
- Implemented the Corto Exchange backend with modules for:
- Configuration:
- Docker support
- Example environment file
- Logging and server configuration
- Documentation:
- Frontend: