S2 StreamStore
S2 is a serverless API and cloud-storage primitive for unlimited, durable, real-time streams. It aims to make data in motion as simple, scalable, replayable, and usage-priced as object storage is for data at rest.
At a Glance
- AI application and agent developers
- Developer tools and real-time application teams
- Observability and telemetry platforms
- Data infrastructure and streaming-pipeline teams
- +2 more
AI Tools by S2 StreamStore
(2)
S2
Serverless Streaming Data Store
Claude Batch Toolkit
Claude Batch Inference Python Library
Discussions
No discussions yet
Be the first to start a discussion about S2 StreamStore
Latest News
s2-lite adds a local web dashboard for browsing basins/streams and using Studio.
Beta Redis Streams API support lets Redis clients append to, replay, and tail S2 streams.
Feldera integration launches with native s2_input and s2_output connectors.
US West and EU North stream locations launch alongside US East.
Products & Services
Managed, serverless durable-stream service. Each stream has a URL and supports append, historical read, live tailing, sequence numbers, replay, concurrency control, access tokens, and bottomless object-storage-backed retention.
Self-hostable single-node S2 server/CLI that can run in memory or against S3-compatible object storage, useful for local development, testing, and private deployments.
REST-first v1 control-plane and data-plane API for basins, streams, records, access tokens, locations, and metrics.
Official SDK support for TypeScript, Python, Go, and Rust, plus CLI access.
Market Position
S2 positions itself as a serverless, durable, high-cardinality alternative to cluster-oriented streaming systems such as Apache Kafka, NATS, Redis Streams, and Amazon Kinesis. Its differentiators are stream-per-session/user cardinality, replayable bottomless retention, HTTP-first simplicity, object-storage economics, and no infrastructure to operate; the company says it is particularly suited to AI-agent and real-time application workloads.
Leadership
Founders
Shikhar Bhushan
Co-founder and builder of S2's serverless-streaming product; his public profile describes his focus as making streams serverless and lists S2.dev and Y Combinator.
Stephen Balogh
Co-founder and CTO. His public professional profile describes prior work as a staff machine-learning engineer in retrieval systems, senior machine-learning engineer in search/matching, and software engineer.
Dwarak Govind Parthiban
Co-founder and founding engineer. His public profile lists prior experience as a Senior Applied Scientist at Etsy and Graduate Research Assistant at the University of Ottawa.
Executive Team
Shikhar Bhushan
Co-founder
Leads the product/company effort around making streams serverless; publicly associated with S2.dev and Y Combinator.
Stephen Balogh
Co-founder and CTO
Engineering leader with prior staff machine-learning, retrieval-systems, search/matching, and software-engineering experience.
Founding Story
The team saw that object storage made static data simple through named resources, usage pricing, durability, and no clusters, but streaming data still required operationally heavy systems. S2 was started to make the stream/log a first-class cloud-storage primitive: records are appended and sequenced durably, can be replayed from the past, and can be tailed live through a simple URL-based API.
Business Model
Revenue Model
Usage-based cloud pricing: customers pay for retained storage, basins, streams, access tokens, data transfers, and API operations. S2 also offers a self-hostable open-source s2-lite path, while managed S2 Cloud is metered hourly and prorated.
Pricing Tiers
Metered hourly and prorated on a 730-hour month.
Data-transfer pricing for writes.
Data-transfer pricing for reads.
Covers operations such as append/read/check-tail, basin management, tokens, metrics, and locations.
Target Markets
- AI application and agent developers
- Developer tools and real-time application teams
- Observability and telemetry platforms
- Data infrastructure and streaming-pipeline teams
- Companies building collaborative, multiplayer, and sync-heavy products
- Teams needing self-hosted or S3-backed streaming infrastructure
- AI agent sessions and agent-to-agent/caller communication
- Token streaming and resumable chat generations
- Telemetry, observability, logs, and metrics
- Event history and replay
- Agent write-ahead logs and crash recovery
- Application and client synchronization
- Kernel
- Trigger.dev
- Beam
- Phonic