v2.36.0
New Features
-
AI Model Enhancements: Introduced a
SupportsEffortLevel
column to the AI Models table, enabling finer control over model capabilities. -
Reasoning Models Integration: Added support for reasoning models, expanding the AI's analytical capabilities.
-
User Context in SkipConversation: Enhanced the
SkipConversation
type by incorporating user information, allowing for more personalized interactions. -
Effort Level Parameter: Introduced an
effortLevel
field toChatParams
, providing more granular control over chat interactions. -
New REST API Endpoints: Added new REST API endpoints to support enhanced functionalities and integrations.
Improvements
-
OpenAI Parameter Update: Replaced the deprecated
max_tokens
parameter withmax_completion_tokens
in OpenAI configurations to align with the latest API standards. -
Anthropic Provider Reasoning: Implemented reasoning capabilities for the Anthropic AI provider, enhancing its inferential responses.
-
Code Generation: Executed code generation processes following the addition of the
SupportsEffortLevel
column to ensure consistency across the codebase.