LanceDB
AI & LLMsDeveloper-friendly OSS embedded retrieval library for multimodal AI. Search More; Manage Less.
Release History
python-v0.33.0-beta.11 fixThis release primarily focuses on fixing an issue related to canonicalizing remote nested field paths.
v0.30.0-beta.11 fixThis release primarily focuses on fixing an issue related to canonicalizing remote nested field paths.
python-v0.33.0-beta.0Breaking7 fixes3 featuresThis release introduces new features across Python and Node.js APIs, including progress reporting and table renaming, while also addressing several bugs related to nested index paths and query routing.
v0.30.0-beta.0Breaking7 fixes3 featuresThis release introduces new features across Python and Node.js APIs, including progress reporting for additions and table renaming, while addressing several bugs related to index path handling and query execution.
v0.29.1-beta.0Breaking2 fixes9 featuresThis release introduces significant feature enhancements across Node.js and Python, including improved namespace management, streaming ingestion support, and better error handling for invalid namespace configurations.
python-v0.32.1-beta.0Breaking2 fixes9 featuresThis release introduces significant feature enhancements across Node.js and Python, including improved namespace management, streaming ingestion support, and query ordering. A key change is stricter validation for namespace modes in Python, which now raises errors instead of failing silently.
python-v0.32.0Breaking13 fixes9 featuresThis release introduces significant refactoring in namespace handling and enterprise integration, alongside new features like Enum support in schema conversion and IVF_HNSW_FLAT index support. Legacy tantivy FTS support has been removed.
v0.29.0Breaking13 fixes9 featuresThis release focuses heavily on namespace operations, FTS enhancements, and PyTorch compatibility improvements, alongside several breaking changes related to namespace naming consolidation.
python-v0.31.0-beta.111 featureThis release introduces a new manifest-enabled directory namespace mode for enhanced data organization.
v0.28.0-beta.111 featureThis release introduces a new manifest-enabled directory namespace mode for enhanced data organization.
python-v0.31.0-beta.102 fixesThis release focuses on security fixes, including enabling hostname verification by default and addressing a cargo-deny advisory. It also removes legacy tantivy FTS support from Python bindings.
v0.28.0-beta.102 fixesThis release focuses on security fixes, including enabling hostname verification by default and addressing a cargo-deny advisory. It also removes legacy tantivy FTS support.
v0.28.0-beta.92 fixesThis release focuses on bug fixes, specifically addressing an issue where the reranker kept processes alive and ensuring namespace tables are materialized correctly upon creation in the Rust backend.
python-v0.31.0-beta.92 fixesThis release focuses on bug fixes, specifically addressing an issue where the reranker kept processes alive and ensuring namespace tables are materialized correctly upon creation in the Rust backend.
v0.28.0-beta.81 fixThis release primarily focuses on a bug fix related to how namespace tables are connected using the Rust backend.
python-v0.31.0-beta.81 fixThis release primarily focuses on a bug fix related to how namespace tables are connected using the Rust backend.
python-v0.31.0-beta.71 fixThis release primarily addresses a bug related to passing the namespace client during schema-only table creation.
v0.28.0-beta.71 fixThis release primarily addresses a bug related to passing the namespace client during schema-only table creation.
python-v0.31.0-beta.61 featureThis release introduces support for nested namespace operations in the Rust backend for database listing and includes documentation fixes.
v0.28.0-beta.61 featureThis release introduces support for nested namespace operations in the Rust backend for database listing and includes documentation fixes.
python-v0.31.0-beta.51 featureThis release introduces support for child namespace operations and JSON serialization capabilities for LanceDBConnection.
v0.28.0-beta.51 featureThis release introduces support for child namespace operations and JSON serialization capabilities for LanceDBConnection.
python-v0.31.0-beta.4v0.28.0-beta.4v0.28.0-beta.3python-v0.31.0-beta.3v0.28.0-beta.23 fixes1 featureThis release introduces a new user_id field in ClientConfig and includes several bug fixes related to vector sanitization and metadata handling, along with migrating the gemini-text provider.
python-v0.31.0-beta.23 fixes1 featureThis release introduces the user_id field to ClientConfig and includes several bug fixes related to vector sanitization, metadata handling, and provider migration.
python-v0.31.0-beta.11 featureThis release introduces support for Enum types during the Pydantic to Arrow schema conversion process.
v0.28.0-beta.11 featureThis release introduces support for Enum types during the Pydantic to Arrow schema conversion process.
python-v0.31.0-beta.0BreakingThis release includes a significant refactor consolidating namespace related naming and enterprise integration.
v0.28.0-beta.0BreakingThis release focuses on internal refactoring, consolidating namespace related naming and enterprise integration, which introduces breaking changes.
python-v0.30.211 fixes6 featuresThis release introduces significant new features like a type-safe expression builder API and support for more vector query types, alongside dependency updates to Lance and various bug fixes for stability and security.
v0.27.211 fixes6 featuresThis release introduces significant new features like a type-safe expression builder in Python and support for more vector types in Node.js, alongside dependency updates and various bug fixes across the codebase.
python-v0.30.2-beta.11 featureThis release primarily updates the internal Lance dependency to version v4.0.0-rc.3.
v0.27.2-beta.11 featureThis release primarily updates the internal Lance dependency to version v4.0.0-rc.3.
v0.27.2-beta.01 fix3 featuresThis release updates the Lance dependency to v3.0.1 and introduces parallel inserts for remote tables. It also adds a progress bar feature for the add() operation.
python-v0.30.2-beta.01 fix3 featuresThis release updates the Lance dependency to v3.0.1 and introduces parallel inserts for remote tables. It also adds a progress bar for the add() operation.
v0.27.11 fixThis release primarily focuses on fixing a bug related to pre-filtering in hybrid search and includes a CI modification to prefer stable releases during checks.
python-v0.30.11 fixThis release primarily focuses on fixing a bug related to pre-filtering in hybrid search and includes a CI modification to prefer stable releases during checks.
v0.27.0Breaking15 fixes21 featuresThis release introduces major features like parallel inserts, expression builder API, and upgrades the core Lance dependency to 3.0.0-rc.3. It also contains breaking changes in Rust APIs for table creation and data addition.
python-v0.30.0Breaking14 fixes21 featuresThis release introduces significant features like expression builder API, parallel inserts, and improved remote table handling, alongside breaking changes in Rust input types for table creation and additions.
v0.27.0-beta.52 featuresThis release introduces the ability to infer JS native arrays and upgrades the underlying Lance library to version 3.0.0-rc.3.
python-v0.30.0-beta.52 featuresThis release introduces the ability to infer JS native arrays and upgrades the underlying Lance library to version 3.0.0-rc.3.
v0.27.0-beta.44 fixes5 featuresThis release introduces several new features, including support for remote index parameters and dictionary-to-SQL struct conversion in updates. It also improves error handling by propagating cast errors during data addition.
python-v0.30.0-beta.43 fixes5 featuresThis release introduces several new features, including support for remote index parameters and dictionary-to-SQL struct conversion in Python updates. Bug fixes address propagated cast errors during additions and improved CI stability.
python-v0.30.0-beta.31 fix1 featureThis release upgrades the Lance dependency to 3.0.0-rc.2, introducing fast_search bindings, and includes a fix to pin pylance for datafusion compatibility.
v0.27.0-beta.31 fix1 featureThis release upgrades the Lance dependency to 3.0.0-rc.2, introducing fast_search bindings, and fixes an issue by pinning the pylance version for datafusion compatibility.
python-v0.30.0-beta.21 fix4 featuresThis release introduces significant performance improvements with parallel inserts and a new expression builder API in Rust, alongside an upgrade to napi-rs for Node.js bindings.
v0.27.0-beta.21 fix4 featuresThis release introduces significant performance improvements with parallel inserts and a new Rust expression builder API, alongside an upgrade to napi-rs for Node.js bindings.
v0.27.0-beta.16 fixes1 featureThis release introduces background checking for dataset updates and includes several bug fixes related to hybrid search, remote table handling, and authentication error retries.
python-v0.30.0-beta.16 fixes1 featureThis release introduces background checking for dataset updates and includes several bug fixes related to hybrid search, remote table reading, and authentication error handling.
v0.27.0-beta.0Breaking3 fixes5 featuresThis release introduces new features like license listing and improved hybrid search explainability, alongside breaking changes in the Rust API for table data ingestion. It also includes several bug fixes and a dependency update to Lance v2.0.1.
python-v0.30.0-beta.0Breaking3 fixes5 featuresThis release introduces breaking changes in Rust bindings to accept RecordBatch types in table creation and addition, alongside several new features like improved hybrid search explanation and PyTorch integration enhancements. Bug fixes address FTS indexing errors and remote table projection.
v0.26.21 fixThis release focuses on build and CI improvements, specifically adding the repository field to package.json for npm provenance in the Node.js CI process.
python-v0.29.21 fixThis release primarily focuses on build and CI improvements, specifically adding the repository field to package.json for npm provenance in the Node.js CI process.
v0.26.1This release focuses on internal refactoring by extracting schema evolution logic and updating the Node.js version used for publishing in the CI pipeline.
python-v0.29.1This release focuses on internal refactoring by extracting schema evolution logic and updating the Node.js version used for publishing in the CI pipeline.
v0.26.010 fixes6 featuresThis release introduces configuration for permutation builder memory limits, adds support for VoyageAI v4 models, and updates the core Lance dependency to v2.0.0. Numerous bug fixes address issues related to Pydantic validation, CI stability, and API behavior.
python-v0.29.010 fixes6 featuresThis release introduces support for VoyageAI v4 models, exposes fast search in the sync API, and updates the core Lance dependency to v2.0.0. Several bug fixes address Pydantic validation, CI stability, and embedding handling.
v0.25.0-beta.06 fixes2 featuresThis release introduces configuration for the permutation builder memory limit and adds support for VoyageAI v4 models. Several bugs related to pydantic validation, PIL usage, and query cancellation have also been resolved.
python-v0.28.0-beta.06 fixes2 featuresThis release introduces support for VoyageAI v4 models and allows configuration of the permutation builder memory limit via an environment variable. Several bug fixes address issues related to Pydantic validation, PIL usage, and query cancellation.
python-v0.27.12 featuresThis release introduces a new feature to check AZURE_STORAGE_ACCOUNT_NAME in remote connections and updates the internal lance dependency to version 1.0.4.
v0.24.11 fix2 featuresThis release introduces a new feature to check AZURE_STORAGE_ACCOUNT_NAME in remote connections and updates the core lance dependency to version 1.0.4. A minor fix was also applied to the Rust MSRV check.
python-v0.27.0Breaking2 fixes6 featuresThis release introduces several performance enhancements like parallelized embedding computations and new model support (voyage-multimodal-3.5). It also includes a breaking change in the Rust implementation by removing default features.
v0.24.0Breaking2 fixes6 featuresThis release introduces several performance enhancements like parallelized embedding computations and adds support for the voyage-multimodal-3.5 model. A significant breaking change involves the removal of default features in the Rust component.
python-v0.27.0-beta.0Breaking2 fixes5 featuresThis release introduces several performance enhancements, including parallelized embedding computations, and adds support for the voyage-multimodal-3.5 model. A breaking change involves the removal of default features in the Rust component.
v0.24.0-beta.0Breaking2 fixes5 featuresThis release introduces several performance and feature enhancements, including support for voyage-multimodal-3.5 and parallelized embedding computations, alongside a breaking change removing default features in the Rust component.
python-v0.26.13 fixesThis release primarily focuses on bug fixes, including updating the HTTP method for namespace description and improving type checking support. Documentation regarding table_names parameters has also been corrected.
v0.23.13 fixesThis release primarily focuses on bug fixes, including updating describe_namespace to use POST and ensuring correct storage option passing for native tables.
v0.23.1-beta.11 fixThis release primarily focuses on a bug fix related to passing namespace storage options providers into native tables.
python-v0.26.1-beta.11 fixThis release primarily focuses on a bug fix related to passing namespace storage options provider into the native table.
python-v0.26.1-beta.01 fixThis release primarily fixes a bug where describe_namespace was not using POST requests, which has now been corrected to allow access to the underlying client.
v0.23.1-beta.01 fixThis release primarily addresses a bug fix by updating describe_namespace to use POST requests, which also allows access to the underlying client.
v0.23.0Breaking4 fixes17 featuresThis release introduces significant features like async namespace support, IVF SQ indexing, and stable row IDs, while also deprecating macOS x86 support and refactoring namespace model usage.
python-v0.26.0Breaking4 fixes17 featuresThis release introduces significant new features, including async namespace support, stable row IDs, and IVF SQ index support, while deprecating macOS x86 support and refactoring namespace handling.
python-v0.26.0-beta.12 featuresThis release introduces new features for the Java SDK, including using the REST namespace and making the client builder generic.
v0.23.0-beta.12 featuresThis release introduces new features for the Java SDK, specifically updating the REST namespace usage and making the client builder generic.
python-v0.26.0-beta.0Breaking1 fix2 featuresThis release introduces server-side namespace querying and new indexing capabilities, alongside critical breaking changes deprecating macOS x86 support and refactoring namespace operations.
v0.23.0-beta.0Breaking1 fix2 featuresThis release introduces server-side namespace querying and new index support (IVF SQ, HNSW aliases), while deprecating macOS x86 support and refactoring namespace model usage.
v0.22.4-beta.32 fixes4 featuresThis release introduces new features like head() for remote tables and support for stable row IDs. It also includes several bug fixes and improvements to documentation and CI infrastructure.
python-v0.25.4-beta.32 fixes4 featuresThis release introduces new features like head() for remote tables and support for stable row IDs. It also includes several bug fixes and improvements to documentation and CI processes.
v0.22.4-beta.23 featuresThis release introduces several new asynchronous features, including support for async namespace connection and async `to_pydantic`, along with an enhancement to merge insert operations.
python-v0.25.4-beta.23 featuresThis release introduces several new asynchronous features, including support for async namespace connection and async `to_pydantic`, along with an enhancement to merge insert operations.
v0.22.4-beta.11 fixThis release primarily focuses on bug fixes, specifically ensuring schema metadata is correctly converted to strings in JsonArrowSchema, alongside several build and CI improvements.
python-v0.25.4-beta.11 fixThis release primarily focuses on bug fixes, specifically ensuring schema metadata is correctly converted to strings in JsonArrowSchema, alongside several build and CI improvements.
python-v0.25.4-beta.01 fix4 featuresThis release introduces features like namespace credentials vending and allows Lance to manage default partition counts. It also includes documentation updates and a dependency bump to Lance version 0.40-0-beta.2.
v0.22.4-beta.01 fix4 featuresThis release introduces features like support for namespace credentials vending and allowing Lance to determine default partition counts. It also includes documentation updates and a dependency bump to Lance version 0.40-0-beta.2.
python-v0.25.31 fix13 featuresThis release introduces significant new features, including IVF_RQ index support, permutation views, and FTS UDTF in SQL, while also updating core dependencies like Lance and removing DynamoDB as a default dependency.
v0.22.31 fix13 featuresThis release introduces significant new features like IVF_RQ index type, permutation views, and FTS UDTF in SQL, alongside dependency updates and removal of the default DynamoDB dependency.
python-v0.25.3-beta.51 featureThis release introduces a new full-text search user-defined table function (UDTF) in SQL.
v0.22.3-beta.51 featureThis release introduces a new feature: adding fts udtf support in SQL queries.
v0.22.3-beta.41 featureThis release introduces an update to the Voyage AI integration and removes outdated documentation regarding the DynamoDB commit store.
python-v0.25.3-beta.41 featureThis release primarily updates the voyage AI integration and removes outdated documentation regarding the DynamoDB commit store.
python-v0.25.3-beta.3No release notes provided.
v0.22.3-beta.3No release notes provided.
python-v0.25.3-beta.21 featureThis release introduces the feature to expose storage options in tables and updates the underlying lance dependency to version v0.38.3-beta.8.
v0.22.3-beta.21 featureThis release introduces the ability to configure storage options directly on tables and updates the underlying lance dependency to a newer beta version.
python-v0.25.3-beta.15 featuresThis release introduces several new features, including an output_schema method for queries and expanded support for multivector models, while also removing the default DynamoDB dependency.
v0.22.3-beta.15 featuresThis release introduces several new features, including an output_schema method for queries and expanded support for multivector models, while also removing the default DynamoDB dependency.
Common Errors
ModuleNotFoundError5 reportsModuleNotFoundError typically arises when a required Python package is not installed in your environment, or when the PYTHONPATH is not correctly set. To fix this, first ensure the missing module (e.g., lance, numpy) is installed using `pip install <module_name>`. If the error persists, verify your Python environment and PYTHONPATH are configured correctly to include the installation path of the missing module.
NotImplementedError3 reportsThe "NotImplementedError" in LanceDB usually arises when attempting to use a feature that is either unavailable in the specific LanceDB deployment (cloud vs. local) or has not yet been implemented in the Python client. To resolve this, confirm the desired feature (e.g., namespaces, certain index types, full-text search) is supported in your LanceDB environment; if not, consider using a supported alternative or check for updates to the LanceDB client library if it's a missing feature in the client. For cloud versions, check the available features supported.
MaxRetryError1 reportMaxRetryError in LanceDB usually arises from network connectivity problems or LanceDB server unavailability, particularly during operations like vector index creation or querying remote tables. Resolve this by ensuring a stable network connection and verifying the LanceDB server is running and accessible. If using cloud providers, check their status and consider implementing retry logic with exponential backoff to gracefully handle temporary failures.
UnboundLocalError1 reportThe UnboundLocalError arises when a variable is referenced before it's assigned within its scope, often inside conditional blocks. To fix this, ensure the variable is always initialized with a default value before the conditional logic is executed; this guarantees it exists regardless of the conditional outcome. For example, initialize `results = None` before the `if` block that might assign actual results.
UnsupportedOperationException1 reportUnsupportedOperationException in LanceDB often arises from trying to perform operations, particularly with UDFs or complex data types like arrays within structs, that are not yet fully implemented or optimized in the underlying data processing engine (Arrow or DataFusion). The fix usually involves either simplifying the query to use supported operations, pre-computing or restructuring the data to avoid the unsupported feature, or updating LanceDB to a version with the necessary functionality implemented, assuming it's a known and resolved issue. If the operation is essential and unfixed, consider filing a bug report detailing your query and data structure for the LanceDB team to address.
ValidationError1 reportValidationError in lancedb often arises from mismatches between the schema defined in your LanceModel (or equivalent data structures) and the actual data being written or queried. Ensure that your data types, field names, and any aliases specified in the LanceModel precisely match the data's structure; double-check for inconsistencies like incorrect data types, misspelled field names, or missing fields. Validate your LanceModel's schema against the expected data format using tools like pydantic's validation features or by inspecting the data before writing to lancedb.
Related AI & LLMs Packages
AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.
Get up and running with OpenAI gpt-oss, DeepSeek-R1, Gemma 3 and other models.
🦜🔗 The platform for reliable agents.
The most powerful and modular diffusion model GUI, api and backend with a graph/nodes interface.
LLM inference in C/C++
GPT4All: Run Local LLMs on Any Device. Open-source and available for commercial use.
Subscribe to Updates
Get notified when new versions are released