Skip to content

Missing ingestion pipeline support for litellm-pgvector #11

Description

@flefevre

Feature request: Add ingestion pipeline support to litellm-pgvector

Description

litellm-pgvector currently does not implement the ingestion pipeline.

This prevents users from using PostgreSQL with pgvector as a complete vector datastore, since documents cannot be processed, embedded, and stored through the standard LiteLLM ingestion workflow.

Expected behavior

litellm-pgvector should support the full ingestion pipeline, including:

  • Document ingestion and chunking
  • Embedding generation
  • Vector and metadata storage in PostgreSQL
  • Compatibility with the existing retrieval workflow

Why this is important

This is a key missing feature for teams that want to use pgvector as their primary vector datastore. Implementing it would unblock end-to-end ingestion and retrieval workflows without requiring a separate vector database or a custom ingestion implementation.

Related discussion: BerriAI/litellm#30878

Activity

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Type

    No type

    Projects

    No projects

      Milestone

      No milestone

      Relationships

      None yet

      Development

      No branches or pull requests

      Issue actions