Skip to main content

Changelog

All notable changes to the LanOnasis platform will be documented here.

[1.2.0] - 2024-01-15​

🚀 Added​

  • Streaming API: Real-time memory retrieval with WebSocket support
  • Hybrid Search: Combined semantic and keyword search for improved accuracy
  • Analytics Dashboard: Usage metrics and insights for monitoring
  • OAuth 2.0 Support: Enterprise SSO integration for secure authentication
  • MCP Integration: Full Model Context Protocol support for AI applications
  • Batch Operations: Process multiple memories efficiently in single requests

🔄 Changed​

  • Improved embedding model to 1536 dimensions for better semantic understanding
  • Reduced search latency by 40% through optimized vector indexing
  • Updated TypeScript SDK to v2.0 with full type safety
  • Enhanced metadata filtering with nested object support
  • Upgraded infrastructure for 10x scale improvements

🐛 Fixed​

  • Memory duplication issue in batch operations
  • Rate limiting edge cases for concurrent requests
  • Metadata filtering with complex nested objects
  • WebSocket connection stability in streaming API
  • Token counting accuracy for large documents

🔒 Security​

  • Added workspace-level encryption for sensitive data
  • Implemented API key rotation mechanism
  • Enhanced audit logging for compliance
  • Added IP allowlisting for enterprise accounts
  • Improved rate limiting algorithms

[1.1.0] - 2023-12-01​

🚀 Added​

  • Python SDK: Official Python client library with full feature parity
  • CLI Tool: Command-line interface for memory management
  • Export/Import: Backup and restore functionality for memories
  • Webhooks: Event notifications for memory operations
  • Custom Embeddings: Support for custom embedding models

🔄 Changed​

  • Increased rate limits for Pro tier users
  • Optimized memory storage for cost efficiency
  • Improved error messages and debugging information
  • Enhanced documentation with more examples

🐛 Fixed​

  • Search relevance scoring inconsistencies
  • Memory update race conditions
  • CLI authentication on Windows systems
  • Pagination issues in large result sets

[1.0.0] - 2023-10-15​

🎉 Initial Release​

  • Core Memory API: Store, retrieve, update, and delete memories
  • Vector Search: Semantic search with state-of-the-art embeddings
  • TypeScript SDK: Full-featured client library for JavaScript/TypeScript
  • REST API: Complete RESTful API with OpenAPI specification
  • Authentication: API key-based authentication with workspace support
  • Rate Limiting: Fair usage policies with tiered limits
  • Documentation: Comprehensive guides and API reference

Key Features​

  • 768-dimension embeddings for semantic search
  • Sub-100ms search latency
  • 99.9% uptime SLA for Pro and Enterprise
  • GDPR compliant data handling
  • Multi-region deployment

Migration Guides​

Upgrading to v1.2.0​

Breaking Changes​

  • search() method now requires explicit namespace parameter
  • Embedding dimensions changed from 768 to 1536
  • Rate limit headers format updated

Migration Steps​

// Old (v1.1.x)
const results = await client.search({
  query: "meeting notes"
})

// New (v1.2.0)
const results = await client.search({
  query: "meeting notes",
  namespace: "default" // Now required
})

Re-indexing for New Embeddings​

# Re-generate embeddings for existing memories
npm run reindex:embeddings -- --model=v2

Upgrading to v1.1.0​

No breaking changes. New features are backward compatible.

Deprecation Notices​

Scheduled for Removal in v2.0.0​

  • Legacy /api/v1/store endpoint (use /api/v1/memory instead)
  • 768-dimension embedding support
  • API key format v1 (migrate to v2 format)

Already Deprecated​

  • client.store() - Use client.upsert() instead
  • includeVector parameter - Use includeEmbedding instead

Upcoming Features (Roadmap)​

Q1 2024​

  • GraphQL API support
  • Real-time collaboration features
  • Advanced analytics dashboard
  • Mobile SDKs (iOS/Android)

Q2 2024​

  • Multi-modal memory support (images, audio)
  • Federated search across workspaces
  • Custom embedding fine-tuning
  • Enterprise data residency options

For questions about migrations or breaking changes, please refer to our Support page or contact your technical account manager.