AIDB 5.0.0 release notes v7
Released: 29 Oct 2025
AIDB 5.0.0 includes the following enhancements and bug fixes:
| Type | Description |
|---|---|
| Enhancement | Generic pipelines: Introduced a new generic pipeline framework (aidb.pipeline) that provides a flexible foundation for building custom AI data processing workflows. |
| Enhancement | PGD support: Added support for EDB Postgres Distributed (PGD), enabling AIDB pipelines and knowledge bases to work in distributed Postgres environments. |
| Enhancement | PostgreSQL 18 support: Added support for PostgreSQL 18. |
| Enhancement | Pipeline state tracking: Added a status field to pipeline metrics to provide better visibility into pipeline execution states. |
| Enhancement | Pipeline auto-processing refinements: Improved the reliability and behavior of automatic pipeline processing, including better handling of pre-existing data. |
| Enhancement | HCP model sync via functions list: Enhanced HCP model synchronization to support syncing models exposed through function lists. |
| Enhancement | NIM NeMo embedding improvements: Updated the input type handling for query and embedding operations on NVIDIA NIM NeMo endpoints. |
| Enhancement | Non-login roles in background workers: Background worker processes can now be run using non-login roles for improved security flexibility. |
| Enhancement | FIPS-compliant TLS: Updated TLS cryptography dependencies to use ring as the default for non-FIPS environments, with continued support for FIPS-compliant builds. |
| Change | Preparer refactored to use generic pipelines: The preparer subsystem has been rebuilt on top of the new generic pipeline framework. Existing preparer configurations remain compatible. |