Pay only for what you use.
Our pricing is simple and transparent.
- Unlimited Datasets
- Storing vectors and metadata in one place
- Out of the box models to vectorise
- Metadata facets & filters
- First class vector support for vector data wrangling and operations
- Customisable Replication and Shards
- Fast vector similarity search
- Flexible vector search (hybrid, multi-vector, chunk-vector, multistep)
- Boost vector search with hybrid approaches (traditional keywords matching, filtering)
- Aggregations, Clustering, Dimensionality reduction, and more
- Global multi-region replication
- Service SLAs
- Isolated environment
- Fine-grained access controls
- Always-on backups
- Dedicated support
|Vector Database Features||Basic||Enterprise|
Real-time & Persistent
Search & Analyse your vectors as soon as it gets inserted in real-time, no index rebuilding required.
Single Store for Metadata & Vectors
Store your metadata and multiple vectors in 1 document and database just like a normal database.
Aggregate vectors in your collection with groupbys to create new collections
Transform vectors in your collection through append, slice and modify
Customisable Replicas & Shards
Customise the amount of shards and replicas for each collection
Facets & Filters
Retrieve facets and filters of your data, to enhance your vector search or analyse subsets of the data
Perform aggregations to understand more about your data and vectors
Vector Similarity Search
Fast nearest neighbours vector similarity search
Combine vectors with traditional keyword matching and boosting
Vector search against chunks (e.g. sentences of paragraphs, images of videos, etc)
Highly Customisable Search
Customise the scoring with different distance metrics, or compare different queries
Clustered search results for diverse range of results
Combine multiple vectors, chunk vectors and traditional for ultimate search
GPU Vector Search
GPU based nearest neighbours for searching through billions and millions of vectors
|Vector Analytics Features||Basic||Enterprise|
Compress, understand and plot your vectors using dimensionality reduction
Clustering & Cluster Aggregation
Group and understand your vectors using clustering and aggregations
Tag your data using vectors from either a user provided tag dictionary or our repository
Deduplicate your data using vectors, on insert or batch
Compare and evaluate the different performance of your vectors
Search History Store
Store and evaluate your search, recommendations, etc.
Vector Embedding Projector
Visualise and interpret your vectors
Encode arrays, dictionaries and fields into vectors
Vector Hub Encoding
Access our massive library of encoders and deep learning embedding models
Premium Vector Hub Encoding
Access our massive library of encoders and deep learning embedding models, created by VecDB
Model Finetuning & Pretraining
Finetune or pretrain models with feedback data or existing metadata using managed pipelines
Dedicated Model Deployment
Deploy dedicated and isolated encoders and vector models for high throughout and zero congestion.
Vectors designed for explainable A.I
Vectors designed for time series data
|Monitoring & Security Features||Basic||Enterprise|
Secure Data Sharing & Replication
Securely share your data with other teams or organisations.
Read Key Provisioning
Provision and deactivate keys for read only access
Fine-grained Access Control
Fine grained access control for different projects, datasets, deployments, etc
Isolated & Dedicated Environment
Dedicated deployment and environment for isolated network, data store and hardware
Multi-region replication for low latency access anywhere in the world
Questions & Answers
VecDB is a vector database for storing, searching, comparing and analysing vectors. Built for machine learning and for the cloud. With VecDB you can quickly prototype and productionise vector based applications like search, recommendations, personalisation, anomaly detection, similarity search, and many many more.
Vector embeddings are meaningful numerical representations of rich data in multi-dimensional space. Vectors can be used to represent any kind of data, such as image, text, audio, videos, users, etc. Once data is represented as vectors it is now possible to accurately search and analyse them using machine learning. It is challenging to develop the resources and infrastructure for generating, searching and analysing vectors. But there’s no escaping the fact: you need vector technology to deliver the best recommendation, discovery, search and translation systems. To learn more: https://getvectorai.com/blog/what-are-vectors-the-silent-disrupter/.
Since VecDB is available through an API it can be ran on any device with an active internet connection. VecDB is a managed database in the cloud so you don't have to maintain, update or scale any service.
For Python we have made a SDK to make the experience as seamless as possible. For example, you can insert your pandas dataframe directly into VecDB with our python SDK. However, if you prefer to interact with the API directly you can still do so.