Deployment Details:
Release Date: November 11, 2025
New Features
Entity Tags System
- New API endpoints for entity tag management (
fetch_entity_tags,fetch_aggregated_entity_tags) - Enhanced data fetching with tag-based filtering and aggregation capabilities
- Improved entity profiling with tag metadata support
Creative Management Enhancements
- New creative endpoints:
GET /creatives/get– Retrieve specific creative by IDPOST /creatives/list– List creatives with pagination and competitor filtering- Enhanced creative processing with better lambda integration
- Improved creative caching and retrieval mechanisms
Database Query Templates
- New query templates for education data, ad group information, and entity tags
- Enhanced UTM tracking with data source and lead source attribution
- Improved entity complete queries with comprehensive field coverage
Improvements
Streaming & Real-time Processing
- Enhanced error handling for streaming operations with
StreamNotFoundexception management - Improved streaming state recovery from Redis with better validation
- Better cancellation handling for streaming jobs
- More robust stream polling with comprehensive error catching
Attachment Processing
- Enhanced image handler with metadata extraction (dimensions, format, file size, aspect ratio)
- Improved attachment parsing with JSON parsing capabilities
- Better error handling for malformed attachment data
- Enhanced vision analysis with metadata preservation
API & Error Handling
- Better exception handlers with proper FastAPI integration
- Improved chat room locking with descriptive error messages (423 status for locked rooms)
- Enhanced billing validation with clearer error messaging
- Better UUID validation across all endpoints
Technical Updates
Dependencies
- Migration to
langchain-tavilyfromlangchain-community.tools.TavilySearchResults - Updated search implementations across onboarding and web search components
- Improved library compatibility and future-proofing
Type Safety & Code Quality
- Centralized UUIDLike type in
Global.Typesmodule - Enhanced type checking with xpytools integration
- Improved import organization and dependency management
- Added PyCharm inspection suppressions for cleaner code analysis
Database & Storage
- Enhanced image attachment storage with metadata fields
- Improved query parameter handling with better type casting
- Better markdown table generation with pipe character escaping
- Enhanced data validation and sanitization
Bug Fixes
- Fixed streaming message serialization with proper type validation and logging
- Resolved attachment processing issues with malformed JSON data
- Fixed brand logo retrieval from brandkit with improved data structure handling
- Corrected variable assignments in ICP generation and data processing
- Enhanced background processing with proper exception handling
- Fixed Redis data parsing in streaming state recovery
- Improved UUID handling in database queries and API endpoints
Code Cleanup
- Removed unused imports and deprecated functions (
safe_cast.py,safe_json_dumper.py) - Added code inspection suppressions for unused class warnings
- Improved logging with better error context and debugging information
- Enhanced documentation with clearer parameter descriptions
- Streamlined exception handling patterns across the codebase
Performance Optimizations
- Better caching for image analysis and creative data
- Improved database query efficiency with optimized templates
- Enhanced streaming performance with better state management
- Reduced redundant API calls through improved caching strategies
Security & Reliability
- Enhanced input validation across API endpoints
- Improved error messaging without exposing sensitive information
- Better exception isolation to prevent cascading failures
- Enhanced data sanitization for user inputs and database queries