EDB components v1.4.0 (LTS)

The EDB folder in Langflow contains components that connect flows to resources managed by Hybrid Manager (HM): knowledge bases, databases, vector indexes, embedding and model server clusters, embedded models, and MCP tooling.

ComponentConnects flow toRequired roles
EDB Knowledge BaseAn AIDB knowledge base in an HM clusterProject Viewer
EDB Vector IndexVector indexes on Postgres tables (pgvector)Project Viewer
EDB DatabaseA Postgres database in an HM clusterProject Viewer
EDB Airman MCPThe Airman MCP server's toolsNone (requires valid PG connection string)
EDB PlatformHM platform MCP tools for the current userDepends on operations invoked
EDB EmbeddingsAn HM-hosted embedding modelAI Model Manager
EDB Model ServerAn HM-hosted inference model (LLM or other)AI Model Manager
EDB Embedded ModelsAIDB embedded models for CPU-based inferenceProject Viewer

See the parent Langflow page for role requirements and Global Variables setup.

EDB Knowledge Base

Langflow component that performs semantic similarity search, optionally combined with column filtering, against an AIDB knowledge base in a Hybrid Manager cluster.

EDB Vector Index

Langflow component that creates, drops, rebuilds, or checks pgvector indexes on PostgreSQL tables in a Hybrid Manager cluster.

EDB Database

Langflow component that produces a Postgres connection URL for a database in a Hybrid Manager cluster.

EDB Airman MCP

Langflow component that exposes a Postgres database to agents as MCP tools, with read-only or unrestricted SQL access.

EDB Platform

Langflow component that exposes Hybrid Manager platform operations (projects, clusters, knowledge bases, and more) as MCP tools to an agent.

EDB Embeddings

Langflow component that produces embedding vectors via a Hybrid Manager-hosted embedding model cluster.

EDB Model Server

Langflow component that calls an inference model (LLM or other) hosted on a Hybrid Manager model server cluster.

EDB Embedded Models

Langflow component that runs CPU-based inference (text and image embeddings, text generation) directly on a Postgres cluster using the aidb extension.