Managed
Vector Search

Filter and rank large sets of high-dimensional vectors
with 10ms search latency on a specialized data infrastructure
Designed for speed, scale and ease of use

Unlock the Value of Unstructured Data

With the help of AI models, all data can be represented as a vector.
Upload the vector embeddings to HIDDB to quickly find similarities within your data set.

Image Search

Search by image and find similar pictures — think shopping for similar looking products. Combine with an NLP model to find images by search term.

Question Answering

Given a new question, e.g. originating from a chat, find semantically matching Q&As from a knowledgebase.

Recommender System

Recommend similar products, blog articles, books, movies — you name it. Display the results in e.g. your webshop or in your CRM.

Identity Verification

Use video, images, or audio to verify a person's identity; or search for similar faces or voices.

Material Research

Search for similar materials and predict material properties.

Audio Search

Search for similar songs, voices, bird sounds, or any other audio.

Quick and Easy Integration into Your Environment

HIDDB's REST API and SDKs for multiple programming languages make it easy and quick to integrate into production applications with just a few lines of code.

Relief Yourself from Managing Operations

Scale up to 10 TB of storage per instance and increase throughput linearly with added replicas - leaving all the intricacies to us:

HIDDB is fully managed and runs in multiple regions with high availability allowing you to focus on your core product.

Answers to Your Queries in the Blink of an Eye

Sub-10ms query latency and high recall rates at scale, even with billions of vectors and tens of thousands of queries per second.

All database instances run in isolation on dedicated VPS so that you don't have to worry about noisy neighbors.

Just Three Steps Away

1
Transform your data

Find an embedding model that works with your type of data. Use one of the many off-the-shelf models available or a self-developed model to transform your data into vector embeddings.

2
Upload your embeddings

Upload your vector embeddings and metadata to HIDDB with the help of our SDKs or REST API. Full CRUD operations with live index updates and guaranteed persistence.

3
Query in real-time

Find nearest neighbors to any vector embedding in < 10 ms with > 95% recall.

Start Free, Then Pay Only
for What You Use

Meet your cost and performance requirements with a simple, usage-based subscription

For small applications and experiments

Single database

Unlimited collections (aka tables)

10 GB of storage

Available locations: EU North (Finnland)

For production applications

Up to 10 databases

Unlimited collections (aka tables)

10 GB of storage per database (support for up to 10 TB is coming soon)

Never sleeps

Available locations: EU Central (Germany) EU North (Finnland)

Up to 10 databases

Unlimited collections (aka tables)

10 GB of storage per database (support for up to 10 TB is coming soon)

Never sleeps

Available locations: EU Central (Germany) EU North (Finnland)

For enterprise applications and custom requirements

Includes all Standard features

Unlimited databases

Data processing pipelines with hosted vector models

Available locations: 12 in the Americas 8 in Europe 3 in Asia 1 in Australia

What Will You Build?

Sign up, and in just a few seconds you'll have a HIDDB instance