Vectorize MongoDB Adapter

Stores and searches Payload embeddings with MongoDB Atlas vector search

Community 98 17/wk MIT v1.1.0

Installation

pnpm add @payloadcms-vectorize/mongodb

About

This adapter lets the Payload vectorize plugin keep its embeddings in MongoDB instead of Postgres. Each knowledge pool gets its own MongoDB collection with a vector search index, which the adapter creates the first time something is written. Semantic search, metadata filtering and the bulk re-embedding flow then run against that collection.

It works with MongoDB Atlas and with self-hosted MongoDB Community 8.2 or newer that has the search engine enabled. Fields used in filters have to be declared up front, and the embedding dimension must match the model. Filtered searches can return fewer results than requested because some filters are applied after the vector lookup.

Package info

Package name
@payloadcms-vectorize/mongodb
Latest version
1.1.0
Unpacked size
95 kB
License
MIT
Weekly downloads
17
Last publish
Jun 22, 2026

More from techiejd

View profile

Similar adapters

More in Database

Database adapter that runs Payload on SurrealDB through its PostgreSQL-compatible SQL interface

Stores Payload data in a Postgres database, with migrations and read replica support

Vercel Postgres database adapter that stores Payload collections and globals in Postgres

Stores Payload data in Cloudflare D1 SQLite databases from Workers

Stores Payload collections and globals in a MongoDB database

Official SQLite database adapter that stores Payload collections and globals in SQLite