NewYour coding agent can read the release notes before it upgrades.Set up the MCP server →
PyPI · #4489 most downloaded on PyPI
An integration package connecting MongoDB and LangChain
Last release 12 days ago
22 Sep 2026
Ships unpredictably
gaps range from 9 days to 7 months
Some releases are documented
notes for 9 of 26 stable releases
Nothing withdrawn
no release was ever pulled
3 years old
28 releases · first in 2024
Nothing published for this version
Fix MongoDBStructuredQueryTranslator to convert LangChain's internal ISO 8601 date/datetime representations ({"date": "YYYY-MM-DD", "type": "date"}) t
Fix MongoDBStructuredQueryTranslator to convert LangChain's internal ISO 8601 date/datetime representations ({"date": "YYYY-MM-DD", "type": "date"}) to Python datetime objects before passing filter values to MongoDB Atlas Vector Search. Previously, these dicts were passed through verbatim, causing an OperationFailure when filtering on date fields.
Add support for auto-embeddings in retrievers:
MongoDBAtlasHybridSearchRetrieverMongoDBAtlasParentDocumentRetrieverMongoDBAtlasSelfQueryRetrieverAdd native reranking ($rerank) support to MongoDBAtlasVectorSearch and all
retrievers. Pass rerank_path, rerank_model, and num_docs_to_rerank to any
search method or retriever to re-rank candidates using the Voyage AI reranker
before returning results. Requires MongoDB Atlas 8.3+ with Native Reranking
enabled in Project Settings. The rerank score is also exposed as rerankScore
in each returned document's metadata.
Note on models (as of public preview): Only rerank-2.5-lite is backed by
real GPUs. Other model names (rerank-2.5, rerank-2, rerank-2-lite) return
a constant score of 0.5987 for all documents and should not be used until
GPU support is confirmed.
One column per month.
Add support for auto-embeddings in MongoDBAtlasVectorSearch.
MongoDBAtlasVectorSearch.Bump langchain-core dependency for CVE-2025-68664.
Add visualization support to MongoDBGraphStore with optional langchain_mongodb[viz].
MongoDBGraphStore with optional langchain_mongodb[viz].search_by_vector method to MongoDBAtlasVectorSearch.Update dependencies to align with LangChain 1.0, which is now the minimum supported version of LangChain. This includes adding langchain_classic as a
langchain_classic as a dependency to support existing Retriever APIs.Pin LangChain version to < 1.0 to avoid compatibility issues while in pre-1.0.
Only list authorized collections when listing collections.
Support separate weights in hybrid retriever.
MongoDBAtlasVectorSearch
if not provided and auto_create_index is set to True.MongoDBAtlasFullTextSearchRetriever.search_field and
MongoDBAtlasVectorSearch.text_key.collection objects consumed by langchain-mongodb.agent_toolkit parser to handle Python and JavaScript objects as part of the command to run.Nothing published for this version
Improve robustness of MongoDBDatabase.run.
MongoDBDatabase.run.Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Nothing published for this version
Your coding agent can read these notes before it upgrades. Set up the MCP server →