orchid-storage-postgres

PostgreSQL storage plugin: chat persistence, checkpointer, and visibility fragment.

orchid-storage-postgres is the PostgreSQL storage plugin for the Orchid AI framework. It provides production-grade PostgreSQL backends for chat persistence, LangGraph checkpointing, and SQL visibility filtering — all in a single installable package that auto-registers via Python entry points.

Installation

pip install orchid-ai orchid-storage-postgres

What It Provides

ComponentClassPurpose
Chat storageOrchidPostgresChatStorageFull OrchidChatStorage implementation backed by asyncpg
Checkpointer_build_postgres_checkpointerLangGraph AsyncPostgresSaver for agent-state persistence
Visibility_build_postgres_filterPostgreSQL $1..$N positional-parameter visibility fragment
MigrationsPostgresMigrationRunnerUnified v001 DDL for all framework-owned tables

Auto-Registration

The package registers itself via importlib.metadata entry points:

[project.entry-points."orchid.visibility_fragments"]
postgres = "orchid_storage_postgres:_register"
[project.entry-points."orchid.checkpointers"]
postgres = "orchid_storage_postgres:_register"

No manual register_visibility_fragment() or register_checkpointer() calls are needed by integrators. The chat storage is referenced via dotted class path in YAML config.