Molecula and Pilosa are now FeatureBase!
An abstract data stack representation

The First OLAP Database Built Entirely on Bitmaps

FeatureBase enables data engineers to process and operate on massive, continuously changing datasets in real time.

About FeatureBase

The secret to FeatureBase’s performance is its efficiency --  we created a new data format that enables the absolute minimum amount of I/O to process analytic queries.

Blazing-Fast Queries
Simple or complex, FeatureBase knocks it out in milliseconds.
Mutable Format
On-the-fly updates and deletes. Operate instantly on your freshest data without the need for preaggregation.
I/O Efficient
Built on bitmaps, FeatureBase offers up to 5-10X reduction in storage footprint and 90% reduction in hardware footprint.
Horizontally Scalable
A distributed, highly-scalable
database built to handle extreme
data volumes.
Secure by Design
FeatureBase is SOC 2 Type II
compliant, and includes features like table-level access control, and query and admin auditing.
Easy to Use
Built-in SQL support plus our
own native query language,
PQL. Automatic decentralized
sharding, and more.

See which option is best for your needs

Tinker and tune FeatureBase yourself with our Open Source offering, or try out our hassle-free, no-maintenance SaaS version of FeatureBase.

Open Source
Join the FeatureBase community and help us build a general-purpose database based on bitmaps.
All the performance and flexibility of FeatureBase without the infrastructure to manage. Try FeatureBase Cloud today.

Why choose FeatureBase?

Looking for an efficient, real-time, mutable database to keep up with your massive-scale analytical workloads, particularly those that require updates and deletes? Look no further -- that's FeatureBase's specialty.

vector image of a trophy
Data format is CPU-friendly resulting in fast performance even on commodity servers
Homomorphic compression + multiple encoding techniques result in cheap in-memory storage
vector image of gear and loading bar
Optimized for analytical queries
Parallel query processing increases performance
Feature-first format (built on bitmaps) reduces I/O to the bare minimum
vector image of profile with binary in it
Built-in SQL support in addition to our native query language, PQL
Connect with applications via rest endpoints for data APIs
Built-in connectors/client libraries for common data sources and languages
magnifying glass over binary
Real-time updates, inserts, and deletes
Multi-value set fields (including time quantums) unlock new use cases

How it works


By collapsing data schemas, FeatureBase eliminates time-consuming preaggregation and helps organizations operate on their freshest data to drive instant decisions. With FeatureBase, organizations are no longer forced to make costly tradeoffs between query speed and ingest volume for concurrent users. Compile, download, or try our cloud version to get started.

Ingest & Explore

FeatureBase is beneficial for organizations that need to access large quantities of real-time data events each day joined with many fragmented, terabyte-scale data sources. FeatureBase excels at complex analytical workloads where source data is fragmented across silos, and where a user or machine wants to apply a number of filters or criteria to a query. See our documentation to get started with ingest.

Connect to application

FeatureBase ingests from a variety of streaming data sources, data warehouses, and change data capture (CDC) plugins. Our getting started guide and community resources will guide you through data modeling, integrations, and other steps needed to connect your products to FeatureBase.

STart free trial

When to Use FeatureBase

thumbs up line-drawing


  • Massive datasets
  • Workloads that require mutability (updates/deletes)
  • Multiple disparate data sources
thumbs down line-drawing


  • Unstructured data with no labels
  • Single record retrieval
  • Data lake functionality


The secret to FeatureBase’s performance is its efficiency --  we created a new data format that enables the absolute minimum amount of I/O to process analytic queries - even on commodity hardware.

Using FeatureBase to analyze activity across our 100M+ monthly login events in real time gives us an entirely new view of our data. Without FeatureBase in our data pipeline, working with this volume of data would be effectively impossible.
Q2 ebanking logo
Adam Blue
CTO, Q2Ebanking
"FeatureBase gives us the ability to analyze significant volumes of both streaming and batch data with extreme efficiency and a minimal hardware footprint.  This has opened possibilities beyond foundational reporting to real-time actionable data to inform operational teams of service disruptions, security issues, and performance anomalies.  All of this to lean forward towards an Adaptive Operational Posture based on real-time infrastructure telemetry."
IT Technology Infrastructure
North Texas Hospital
FeatureBase is hands-down a game-changer for our company and clients. The technology behind the blistering performance speeds is unparalleled and couldn't be backed by a better team. Our clients were experiencing query result times exceeding 30 seconds and now with FeatureBase the same queries take less than 100ms.
Pumpjack Dataworks logo
Corey Zwart
Head of Engineering, Pumpjack Dataworks
Moodmack Power and Renewables are embarking on bringing an industry-first, high granular level of insight into a traditionally opaque market. FeatureBase allows us to analyze significant volumes of both streaming and batch data, which was not possible prior. In addition, with the hyper-increased exposure and access to our vast data reserves, we can unlock future products and services that were thought not possible.
Wood Mackenzie logo
Matthew Boyda
Vice President, P&R/Power Solutions, Wood Mackenzie
Trusted by companies like:

Get Started for Free

Open source install commands are included below.


git clone

cd featurebase

make install

featurebase server --handler.allowed-origins=http://localhost:3000

git clone

cd featurebase

make install

featurebase server --handler.allowed-origins=http://localhost:3000