get started


Collect and store real-time data streams. Keep data forever with infinite retention. Slice and condense data along any dimension with hyper-indexing.

  • Zero retention limits.
  • Zero servers to maintain.
  • Zero vendor lock-in.
durable streams
with infinite retention
fine-grained seeking
across full history
hyper-index with
many, precise partitions
to Amazon S3

Create an integration in minutes with the Pyroclast CLI and our open-source SDKs.

$ pyroctl inspect topic my-event-stream
       21.6% hot

                    hot       archive
data volume         43.2 gb   156.6 gb
avg. read latency   3 ms      54 ms
avg. write latency  5 ms      67 ms
optimistic cache    -         640 mb

Introductory pricing for pyro/store


per partition
per hour


per 1,000,000
payload units



Free tier

3 partitions


No credit card required.

A new retention paradigm

Never worry about deleted live data. Eliminate merging and provenance headaches when mixing new and old data.

pyro/store automatically archives and restores data to and from inexpensive storage, presenting a unified stream of all your data on demand.


Turn your data set into ultra-small, focused views—based on any dimension of your data.

pyro/store empowers you to reap the benefits of a many-partition access pattern, without the operational concerns inherent in other log architectures.

Full-breadth & laser-focused

Rapidly navigate to the exact location in your data stream that you care about.

Seek by time, message count, partition, and more across your entire data history, irrespective of size.

Seamless archival

Keep your data with you. pyro/store manages data at rest in a storage location you control, in a format that you can read without any vendor-specific tools.

limited release


Analyze, detect and react to event data.

Quickly create powerful processing workflows with built-in streaming primitives, or create your own.

Real-time queries
Streaming ETL
Pattern matching
Reactive triggering
Interactive simulation
Declarative YAML design