---
layout: article
title: VectorsDB
description: Store vector embeddings and run similarity search with Appwrite VectorsDB to power semantic search, recommendations, and other AI features.
---

Appwrite VectorsDB lets you store vector embeddings and run similarity search over them.
A collection is created with a fixed `dimension`, every document holds an `embeddings` vector of that length plus optional `metadata`, and an HNSW index keeps similarity search fast as your data grows.

**Looking for file storage?**

Databases store data, if you need to store files like images, PDFs or videos, use [Appwrite Storage](/docs/products/storage).

You organize data into databases, collections, and documents, the same way you do across Appwrite Databases. What sets VectorsDB apart is the fixed schema built for vectors and the ability to generate text embeddings and search by similarity.

[Quick start](/docs/products/databases/vectorsdb/quick-start)

# Key concepts

- [Databases](/docs/products/databases/vectorsdb/databases): Create and manage vector databases, the largest organizational unit.
- [Collections](/docs/products/databases/vectorsdb/collections): Created with a required `dimension`, the length of the vectors they hold, plus an optional `metadata` object per document.
- [Documents](/docs/products/databases/vectorsdb/documents): Store one embedding under `embeddings` and any associated JSON under `metadata`.
- [Embeddings](/docs/products/databases/vectorsdb/embeddings): Generate embeddings from text with built-in models, without a separate embedding service.
- [Permissions](/docs/products/databases/vectorsdb/permissions): Control access at the collection and document level.
- [Queries](/docs/products/databases/vectorsdb/queries): Filter documents by their `metadata` with comparison, string, and logical operators.
- [Order](/docs/products/databases/vectorsdb/order): Sort results by metadata fields.
- [Backups](/docs/products/databases/vectorsdb/backups): Schedule backup policies and restore from any backup.

# Guides

- [Vector search](/docs/products/databases/vectorsdb/vector-search): Create an HNSW index on `embeddings`, then rank documents by cosine, dot product, or Euclidean distance.
- [Pagination](/docs/products/databases/vectorsdb/pagination): Page through results with offset or cursor pagination.
- [Transactions](/docs/products/databases/vectorsdb/transactions): Stage operations and commit or roll them back atomically.
- [Bulk operations](/docs/products/databases/vectorsdb/bulk-operations): Create, update, upsert, or delete many documents in one request.
- [Timestamp overrides](/docs/products/databases/vectorsdb/timestamp-overrides): Set custom creation and update timestamps during imports.
- [CSV imports](/docs/products/databases/vectorsdb/csv-imports): Import documents from CSV files.
- [CSV exports](/docs/products/databases/vectorsdb/csv-exports): Export collections to CSV files.