- NPX
- Docker
- feat: added http handlers for image generation endpoints
- feat: improved model validation for provider-prefixed model configurations
- fix: added support for model lookup in Google GenAI integration by path parameter (fixes using VK provider routing for GenAI integration)
- chore: added case-insensitive helper methods for header and query parameter lookups in HTTPRequest
- feat: add raw request data to bifrost error responses
- fix: add support for AdditionalProperties structures (both boolean and object types)
- fix: improve thought signature handling in gemini for function calls
- fix: enhance citations structure to support multiple citation types
- fix: anthropic streaming events through integration
- feat: adds env variable indicators on UI
- feat: introduces new EnvVar for env backed config fields
- fix: missing request type in error response for anthropic SDK integration
- feat: added support for web search tools in Openai, Anthropic and Gemini
- fix: turn JSON array tool result into JSON object for Bedrock Converse API - @Hieu Hoang
- fix: remove the configuration to clear usages on update when new max limit < current usage
- feat: added image generation request and response support
- chore: added case-insensitive helper methods for header and query parameter lookups in HTTPRequest
- feat: added support for path parameter lookups in HTTPRequest
- fix: missing request type in error response for anthropic SDK integration
- feat: add raw request data to bifrost error responses
- fix: add support for AdditionalProperties structures (both boolean and object types)
- fix: improve thought signature handling in Gemini for function calls
- fix: enhance citations structure to support multiple citation types
- fix: anthropic streaming events through integration
- feat: added support for web search tools in OpenAI, Anthropic and Gemini
- fix: turn JSON array tool result into JSON object for Bedrock Converse API
- feat: add image generation streaming accumulation support
- feat: Improved model matching to support provider-prefixed model names (e.g., “openai/gpt-4”)
- feat: Adds rdb backed distributed locks
- feat: fixed weighted provider routing to correctly match provider-prefixed models in allowed lists
- fix: added support for model lookup in Google GenAI integration by path parameter
- chore: updated core to v1.3.9 and framework to v1.2.10
- fix: remove the configuration to clear usages on update when new max limit < current usage
- chore: updated core to v1.3.9 and framework to v1.2.10
- chore: updated core to v1.3.9 and framework to v1.2.10
- chore: updated core to v1.3.9 and framework to v1.2.9
- chore: updated core to v1.3.9 and framework to v1.2.10
- chore: updated core to v1.3.9 and framework to v1.2.10
- feat: added semantic caching support for image generation
- chore: updated core to v1.3.9 and framework to v1.2.10
- chore: updated core to v1.3.9 and framework to v1.2.10

