Postgres Memory+ icon

Postgres Memory+

Stores the chat history in Postgres table with schema support.

Postgres Memory+

AI-generated

Summary

Store and retrieve chat history and session metadata in a PostgreSQL database with optional semantic search and working memory features.

Inputs

  • Session ID (required) — Identifies the chat session to store or retrieve messages for. Can be sourced from a connected Chat Trigger node or defined via expression or static text.
  • Schema Name — The PostgreSQL schema in which to store the chat history table; will be created automatically if it does not exist.
  • Table Name — The name of the table to store chat histories; created if it does not exist.
  • Options — Collection of configurable options including enabling semantic search, session tracking, working memory, message range for context, number of top semantic search results, user ID for scoping, and whether to store full message metadata.
  • Vector Store Input — Required if semantic search is enabled; connects to a Vector Store node providing embeddings for semantic similarity searches.

Output shape

AI Memory object representing the chat history and optionally working memory, supporting memory variable loading and context saving.

If semantic search is enabled, the node consumes a Vector Store input for embedding queries. It automatically creates necessary tables/schemas as needed and can track sessions and user-scoped working memory. Stored messages can be sanitized to omit intermediate metadata unless configured otherwise.

Links

Discussion