​
August 2025

​
Version 3.2.5 - August 1, 2025

🔍 Search & Performance

  • Query Caching Optimization: 40% faster response times for repeated queries
  • Index Compaction: Reduced storage footprint by 25% through better compression
  • Load Balancer Improvements: Better traffic distribution across nodes
  • Connection Pool Optimization: Reduced connection overhead by 30%

​
Version 3.2.4 - July 25, 2025

📊 Analytics & Quality

  • Query Performance Metrics: Detailed latency breakdown per query stage
  • Usage Analytics Dashboard: Real-time monitoring of API usage patterns
  • Relevance Scoring Improvements: Better ranking for domain-specific queries
  • Duplicate Detection: Automatic deduplication of similar search results

​
Version 3.2.3 - July 20, 2025

⚡ Vector Operations & Security

  • Vector Quantization: 8-bit quantization reduces memory usage by 50%
  • Batch Processing: Support for processing up to 500 documents simultaneously
  • Rate Limiting: Per-tenant rate limiting with burst allowance
  • Input Validation: Enhanced validation for uploaded document formats

​
Version 3.2.2 - July 15, 2025

🔧 API & Search Features

  • Timeout Configuration: Configurable timeout settings for long-running queries
  • Error Handling: More descriptive error messages with suggested fixes
  • Fuzzy Matching: Typo-tolerant search with configurable edit distance
  • Filter Optimization: 60% faster filtering on metadata fields

​
Version 3.2.1 - July 10, 2025

🧠 AI & Scalability

  • Context Window Expansion: Support for 16K token context in answer generation
  • Citation Accuracy: Improved precision in source attribution
  • Horizontal Scaling: Auto-scaling based on query load
  • Read Replicas: Read-only replicas for better query distribution

​
Version 3.2.0 - July 5, 2025

🎛️ Configuration & Processing

  • Dynamic Configuration: Update search parameters without restart
  • Environment Variables: Better configuration management through env vars
  • Incremental Indexing: Update indices without full rebuild
  • Background Processing: Non-blocking document processing pipeline

​
June 2025

​
Version 3.1.6 - June 28, 2025

📊 Monitoring & Vector Operations

  • Health Checks: Comprehensive health monitoring endpoints
  • Metrics Export: Prometheus-compatible metrics for monitoring
  • HNSW Index Optimization: 30% faster nearest neighbor search
  • Embedding Model Updates: Support for latest sentence-transformer models

​
Version 3.1.5 - June 21, 2025

🛠️ Developer Experience & Content

  • SDK Updates: Improved Python and JavaScript SDK with better error handling
  • API Documentation: Interactive examples with curl and code samples
  • PDF Layout Detection: Better extraction from complex PDF layouts
  • Table Extraction: Improved parsing of tabular data from documents

​
Version 3.1.4 - June 14, 2025

🎯 Search Precision & Storage

  • Semantic Search Tuning: Fine-tuned embeddings for technical content
  • Query Expansion: Automatic expansion of short queries with related terms
  • Compression Improvements: Lossless compression reduces storage by 35%
  • Garbage Collection: Automatic cleanup of unused vector embeddings

​
Version 3.1.3 - June 7, 2025

🚀 Performance & Bug Fixes

  • Multi-Threading: Parallel processing for batch document uploads
  • Memory Management: Optimized memory usage for large document collections
  • Memory Leak Fix: Resolved memory leaks in long-running search operations
  • Unicode Handling: Better support for non-Latin character sets

​
Version 3.1.2 - May 31, 2025

📱 Multi-Modal & Customization

  • Image Embedding: Support for image-text multi-modal search
  • OCR Integration: Extract and index text from images in documents
  • Custom Scoring: User-defined scoring functions for specialized domains
  • Boost Parameters: Configurable boost factors for recent or popular content

​
Version 3.1.1 - May 24, 2025

🔄 Real-time & Analytics

  • Streaming Indexing: Real-time indexing of new documents
  • Live Configuration: Update search parameters without system restart
  • Query Analytics: Track popular queries and search patterns
  • Performance Insights: Identify slow queries and optimization opportunities

​
Version 3.1.0 - May 17, 2025

🎯 Answer Quality & Search

  • Response Filtering: Filter out low-confidence answers automatically
  • Source Diversity: Ensure answers draw from multiple diverse sources
  • Hybrid Search Improvements: Better fusion of semantic and keyword search
  • Personalization: User-specific search result ranking based on history

​
Version 3.0.9 - May 10, 2025

⚡ Optimization & Reliability

  • Query Optimization: 25% reduction in average query latency
  • Connection Pooling: Improved database connection management
  • Circuit Breaker: Automatic failover for degraded services
  • Retry Logic: Intelligent retry mechanisms for transient failures

​
Version 3.0.8 - May 3, 2025

🎉 Major Release & Data Management

  • Agent Routing: Automatic routing to specialized AI agents
  • Multi-Step Reasoning: Break complex queries into logical steps
  • Data Retention Policies: Configurable retention for different content types
  • Backup Optimization: Incremental backups with faster restore times

​
April 2025

​
Version 2.9.7 - April 26, 2025

🔧 API & Search Accuracy

  • Batch Operations: Bulk delete and update operations
  • Webhook Support: Real-time notifications for document processing events
  • Relevance Tuning: Improved relevance scoring for technical documentation
  • Noise Reduction: Better filtering of irrelevant search results

​
Version 2.9.6 - April 19, 2025

💾 Storage & Processing

  • Index Optimization: 20% reduction in index size through better encoding
  • Cold Storage: Automatic archival of rarely accessed content
  • Parallel Processing: Multi-threaded document processing
  • Queue Management: Better handling of large document upload queues

​
Version 2.9.5 - April 12, 2025

🛠️ Infrastructure & Content

  • Auto-Scaling: Automatic scaling based on traffic patterns
  • Load Distribution: Improved load balancing across service instances
  • Improved OCR: Better text extraction from scanned documents
  • Metadata Enrichment: Automatic extraction of document metadata

​
Version 2.9.4 - April 5, 2025

🎛️ UX & Performance

  • Response Streaming: Real-time streaming of search results
  • Progress Indicators: Upload and processing progress tracking
  • Vector Index Optimization: 40% faster similarity search
  • Memory Usage: Reduced memory footprint for large indices

​
March 2025

​
Version 2.8.8 - March 29, 2025

🔍 Query & Monitoring

  • Query Normalization: Better handling of diverse query formats
  • Spell Correction: Automatic correction of typos in search queries
  • Detailed Logging: Enhanced logging for debugging and optimization
  • Performance Metrics: Real-time performance monitoring dashboard

​
Version 2.8.7 - March 22, 2025

🛡️ Security & Quality

  • API Key Rotation: Support for seamless API key rotation
  • Request Validation: Enhanced input validation and sanitization
  • Duplicate Filtering: Automatic removal of near-duplicate results
  • Content Freshness: Preference scoring for recently updated content

​
Version 2.8.6 - March 15, 2025

⚡ Speed & Developer Tools

  • Cache Optimization: Smarter caching strategies for frequent queries
  • Index Compression: Reduced index size with maintained accuracy
  • Debug Mode: Detailed query execution information for troubleshooting
  • API Testing: Built-in API testing interface for development

​
Version 2.8.5 - March 8, 2025

📈 Scaling & Answer Generation

  • Connection Pooling: Optimized database connection management
  • Resource Allocation: Dynamic resource allocation based on workload
  • Response Formatting: Better structured responses with bullet points
  • Context Preservation: Improved context retention in longer conversations

​
Version 2.8.4 - March 1, 2025

🔍 Search & AI Improvements

  • Faceted Search: Multi-dimensional filtering capabilities
  • Search History: Track and reuse previous successful searches
  • Model Updates: Updated to latest embedding models for better accuracy
  • Fine-tuning: Domain-specific fine-tuning for technical content

​
February 2025

​
Version 2.2.0 - February 28, 2025

🎯 New Features

  • Advanced Analytics Dashboard: Real-time insights into query patterns and user behavior
  • Custom Model Support: Integration with custom fine-tuned models for specialized domains
  • Workflow Automation: Automated content processing pipelines with custom triggers
  • Vector Similarity Tuning: Fine-grained control over similarity thresholds and scoring algorithms
  • Multi-Modal Search: Support for image, audio, and video embeddings alongside text
  • Federated Search: Query across multiple vector databases simultaneously
  • Creative Search Mode: Generate creative and imaginative answers beyond factual content
  • Web Search Integration: Real-time web search capabilities with search_modes configuration

⚡ Performance & Scale

  • Lightning-fast Indexing: 60% improvement in document processing speed with parallel processing
  • Smart Caching: Intelligent cache invalidation reduces API response times by 75%
  • Optimized Memory Usage: Reduced memory footprint for large-scale deployments
  • Auto-scaling Infrastructure: Dynamic scaling based on query load and index size
  • Distributed Vector Storage: Sharded vector indices for billions of embeddings
  • Query Optimization: Advanced query planning and execution for complex searches

🛡️ Quality of Service

  • Enterprise-grade Reliability: Automatic failover and high availability
  • Real-time Health Monitoring: Comprehensive metrics and alerting system
  • Circuit Breaker Pattern: Automatic degradation and recovery mechanisms
  • Load Balancing: Intelligent request distribution across multiple regions
  • Data Consistency: Strong consistency with eventual consistency options

​
Version 2.1.5 - February 14, 2025

💝 Valentine's Update

  • Enhanced User Experience: Streamlined onboarding flow for new developers
  • API Rate Limiting: Flexible rate limiting with burst allowances
  • Multi-language Support: Expanded language detection and processing capabilities
  • Approximate Nearest Neighbor (ANN): HNSW algorithm implementation for faster similarity search
  • Vector Compression: Quantization techniques reducing storage by 70% with minimal accuracy loss
  • Batch Query Processing: Execute up to 1000 queries in a single API call
  • Streaming Responses: Real-time streaming of answers with stream flag for better UX
  • Auto Agent Routing: Intelligent routing to best-suited AI agents with auto_agent_routing

🔍 Search Enhancements

  • Hybrid Search: Combine dense vector search with sparse keyword matching using search_alpha
  • Semantic Filtering: Filter results based on semantic similarity before ranking
  • Query Expansion: Automatic query enhancement using related terms and concepts
  • Result Diversification: Ensure diverse results to avoid echo chambers
  • Recency Bias Control: Fine-tune preference for recent content with recency_bias parameter
  • Context-Aware Search: Restrict search to specific documents using context_list

​
Version 2.1.2 - February 7, 2025

🔧 Maintenance Release

  • Security Patches: Updated dependencies and security improvements
  • Error Handling: Better error messages with actionable suggestions
  • Documentation Updates: Enhanced code examples and integration guides
  • Index Optimization: Automatic background index compaction and cleanup
  • Memory Leak Fixes: Resolved memory leaks in long-running search operations
  • Connection Pooling: Improved database connection management for high-throughput scenarios

📊 Scalability Improvements

  • Horizontal Partitioning: Distribute large indices across multiple nodes
  • Read Replicas: Support for read-only replicas to handle increased query load
  • Asynchronous Indexing: Non-blocking index updates for better write performance
  • Resource Throttling: Intelligent resource allocation based on workload patterns

​
January 2025

​
Version 2.1.0 - January 23, 2025

🚀 New Features

  • Enhanced User Memory System: Added generate-user-memory endpoint for automatic memory generation
  • Improved Batch Upload: New batch processing capabilities with better error handling
  • App Integration: Native support for app-specific data sources with app-upload endpoint
  • Vector Index Versioning: Maintain multiple versions of indices for A/B testing
  • Real-time Streaming Search: WebSocket-based streaming search results
  • Custom Distance Metrics: Support for cosine, euclidean, dot product, and custom similarity functions
  • Multi-Step Reasoning Engine: Break down complex queries with multi_step_reasoning flag
  • Source Highlighting: Precise content location with highlight_chunks and layout coordinates
  • User Personalization: Custom user instructions and personalized responses with user_name

⚡ Performance & Scale

  • Query Performance: 40% faster response times for search and QnA endpoints
  • Memory Retrieval: Enhanced personalization with improved ranking algorithms
  • API Documentation: Complete overhaul of API reference with interactive examples
  • Massive Scale Support: Handle up to 100M+ vectors with sub-100ms query latency
  • GPU Acceleration: CUDA-optimized vector operations for 10x faster indexing
  • Adaptive Indexing: Dynamic index structure optimization based on query patterns

🔧 Quality & Reliability

  • Fixed memory deletion edge cases in multi-tenant environments
  • Resolved webpage scraping timeout issues for large documents
  • Improved error handling for malformed document uploads
  • Data Validation: Comprehensive input validation and sanitization
  • Backup & Recovery: Automated incremental backups with point-in-time recovery
  • Monitoring & Alerting: Proactive issue detection with detailed performance metrics

​
Version 2.0.5 - January 18, 2025

🔧 Updates

  • Text Upload Optimization: Renamed and optimized markdown upload to text upload for better clarity
  • Search Endpoint Enhancement: Improved search relevance scoring with machine learning models
  • Documentation Updates: Enhanced use cases and implementation guides
  • Vector Quantization: Product quantization for 8x storage reduction
  • Hot/Cold Storage: Automatic tiering of frequently vs rarely accessed vectors
  • Query Result Caching: Redis-based caching for frequently requested searches
  • AI Generation Control: Toggle AI-powered responses with ai_generation flag
  • Metadata Filtering: Advanced document filtering using source_title and source_type

🎯 Search Quality

  • Precision Tuning: Fine-tuned retrieval models for domain-specific content
  • Recall Optimization: Improved coverage for long-tail and niche queries
  • Relevance Scoring: Advanced BM25 + vector hybrid scoring algorithms
  • Query Understanding: Better handling of ambiguous and multi-intent queries
  • Result Count Control: Configurable result limits with top_n and max_chunks parameters

​
January 2025

​
Version 2.0.0 - January 9, 2025

🎉 Major Release

  • Multi-Step Reasoning: Revolutionary multi-step query processing for complex questions
  • Self-Improving System: Automatic learning from user interactions and feedback
  • Advanced Personalization: Deep user profiling with preference learning
  • Web Search Integration: Real-time web search capabilities for up-to-date information
  • Enterprise Vector Database: Full-featured vector database with advanced transaction support
  • Multi-Region Deployment: Global deployment with automatic geo-routing

🚀 Scale & Performance

  • Petabyte Scale: Support for petabyte-scale vector collections
  • Million QPS: Handle millions of queries per second with linear scaling
  • Sub-millisecond Latency: P99 latency under 5ms for most queries
  • Zero-Downtime Updates: Live index updates without service interruption
  • Automatic Sharding: Intelligent data distribution across clusters

🔄 Breaking Changes

  • Updated API versioning scheme (v1 → v2)
  • Modified response format for search endpoints
  • Renamed several endpoint parameters for consistency
  • High Availability: 99.95% uptime with automated monitoring

​
December 2024

​
Version 1.8.0 - December 31, 2024

✨ New Features

  • Custom Metadata Support: Rich metadata tagging with tenant_metadata and document_metadata
  • Multi-Tenant Architecture: Complete isolation with tenant_id and sub_tenant_id support
  • Advanced Filtering: Powerful filtering options for knowledge retrieval
  • Batch Document Processing: Upload multiple documents with batch_upload endpoint
  • Session Management: Conversation tracking and context continuity with session_id
  • Vector Similarity Search: Advanced ANN algorithms (HNSW, IVF, LSH)
  • Incremental Index Updates: Update vectors without full reindexing
  • Distributed Computing: Spark integration for large-scale batch operations

📈 Enterprise Scale

  • High Availability: Multi-master setup with automatic failover
  • Disaster Recovery: Cross-region replication with RPO < 1 minute
  • Security Compliance: SOC2, GDPR, HIPAA compliance certifications
  • Audit Logging: Comprehensive audit trails for all operations

​
Version 1.7.2 - December 17, 2024

🛡️ Security & Performance

  • Enhanced security protocols for API authentication
  • Improved rate limiting and request validation
  • Optimized vector indexing for faster retrieval
  • End-to-End Encryption: AES-256 encryption for vectors at rest and in transit
  • Performance Monitoring: Real-time performance dashboards and SLI tracking
  • Cost Optimization: Intelligent resource allocation reducing costs by 35%

🔍 Advanced Search

  • Semantic Search Refinement: Improved understanding of user intent
  • Faceted Search: Multi-dimensional filtering and faceting capabilities
  • Search Analytics: Detailed query performance and result quality metrics
  • Fuzzy Matching: Typo-tolerant search with configurable edit distance

​
November 2024

​
Version 1.7.0 - November 28, 2024

🔍 Search & Memory

  • Contextual Answers: AI-powered answers with source attribution
  • Memory Continuity: Persistent memory across user sessions
  • Document Processing: Enhanced support for multiple file formats

​
October 2024

​
Version 1.6.0 - October 25, 2024

🚀 Platform Launch

  • Initial Public Release: Cortex SDK goes live with core functionality
  • Knowledge Ingestion: Document upload and webpage scraping
  • Query & Retrieval: Semantic search and question-answering
  • User Memory System: Basic user memory management
  • API Documentation: Comprehensive developer documentation

​
Stay Updated

Get Support

Have questions about an update? Reach out to our team for help.

Feature Requests

Share your ideas for new features and improvements.
Migration Guide: When updating to a new major version, check our migration guides in the documentation to ensure smooth transitions.
Stay Informed: Subscribe to our updates by contacting our team to be notified of new releases and important changes.