github mem0ai/mem0 v1.0.0

3 days ago

πŸš€ What’s New in Mem0 v1.0.0

This release marks a major milestone for Mem0 β€” introducing v1.0.0 with key architectural improvements, expanded vector store integrations, and broader LLM provider compatibility. Expect better performance, more stability, and enhanced developer experience across SDKs.

🌟 New Features

  • Added Azure MySQL support for improved cloud database flexibility by @parshvadaftari
  • Introduced Azure AI Search Vector Store support for the TypeScript SDK by @G26karthik
  • Added Tool Call support for LangchainLLM by @soapun
  • Enabled custom model and parameters for Hugging Face with huggingface_base_url by @vedant381
  • Added default LLM configuration updates for streamlined setup by @parshvadaftari
  • Enabled async mode as default for optimized performance by @kk2211
  • Reranker support - cohere, zeroentropy, huggingface, sentence transformers and LLMs by @parshvadaftari
  • Added metadata filtering for the OSS @parshvadaftari
  • Added Assisstant memory retrieval by @parshvadaftari
  • Improved prompt for better memory retrieval by @parshvadaftari

🧠 Improvements & Fixes

  • Refactored v2 search and update examples by @parshvadaftari
  • Fixed Databricks Vector Store integration by @hayescode
  • Fixed Milvus DB bug and added test coverage by @parshvadaftari
  • Fixed Weaviate search method by @matanco1
  • Fixed non-serializable objects in config deepcopy by @lsvishaal
  • Fixed bug with thinking LLM in vLLM by @dog-last
  • Fixed license and README metadata by @lsvishaal
  • Fixed hardcoded embedding dimensions and TypeScript build issues by @Mrinank-Bhowmick
  • Fixed missing β€˜data’ key handling during search operations by @lsvishaal
  • Improved Get Memories playground and documentation examples by @parshvadaftari
  • Added detailed explanation for output_format v1.1 parameter by @lsvishaal
  • Updated dependency compatibility with OpenAI 2.x by @yashikabadaya
  • Updated main script and removed stale code by @deshraj
  • Added version=v2 as default param in AI SDK add calls by @whysosaket
  • Validated embedding_dims for Kuzu integration by @soapun

🧩 Documentation Updates

πŸ‘₯ New Contributors

Full Changelog: [v0.1.118 β†’ v1.0.0](v0.1.118...v1.0.0)

Don't miss a new mem0 release

NewReleases is sending notifications on new releases.