# StreamNative > StreamNative is the Streamhouse for the age of AI. The StreamNative Data Platform is a shared data architecture that captures, transports, transforms, governs, and serves the current state of a business continuously, so applications, analytics, and AI agents can act on it — real-time, production-native, and decentralized. It runs native Apache Kafka and Apache Pulsar on Lakestream, the open stream storage, with Ursa as its implementation, and keeps every interface open: Kafka and Pulsar for streams, Apache Iceberg and Delta Lake for tables, PostgreSQL-compatible SQL for queries, and MCP for agents. Founded by the original creators of Apache Pulsar at Yahoo, StreamNative serves global enterprises across financial services, technology, retail, logistics, and more. ## Core Products ### StreamNative Data Platform - URL: https://streamnative.io/data-platform/overview - Category: Streamhouse — Real-Time Data Platform The StreamNative Data Platform is the Streamhouse for apps, analytics, and agents: a shared data architecture for the age of AI that keeps the current state of the business continuously available. Streamhouse and lakehouse architectures are built on the same open infrastructure — object storage, open table formats, and catalogs. The lakehouse is data to analyze the business; the Streamhouse is data to run it. **How it works:** - Capture: Universal Connect runs Kafka Connect and Pulsar IO connectors, including change data capture, on one runtime. Universal Linking (Public Preview) mirrors data and schemas from existing Kafka clusters without downtime. - Transport: native Apache Kafka and Apache Pulsar run on Ursa, over Lakestream, the open stream storage. The protocol is a choice of interface, not a choice of data silo. - Transform and serve: SQL Workspace (Private Preview, powered by RisingWave) keeps materialized views over streams continuously current behind a PostgreSQL-compatible endpoint. Apache Flink, powered by Ververica, is in Early Access. - Govern: streams land as Apache Iceberg and Delta Lake tables and register in the catalogs you already run, including Databricks Unity Catalog, Snowflake Horizon Catalog, and Iceberg REST catalogs. **Key attributes:** - Real-time: current to the millisecond rather than hours old - Production-native: engineered to production service levels for applications, analytics, and agents - Decentralized: meets data where it already lives rather than consolidating it first - Open interfaces: Kafka and Pulsar for streams, Iceberg and Delta for tables, PostgreSQL for queries, and the StreamNative MCP Server for agents ### Kafka Service - URL: https://streamnative.io/data-streaming/kafka - Category: Real-time Data Streaming StreamNative provides a native Kafka service — not a compatibility layer or protocol translation — powered by Ursa Engine. Organizations keep their existing Kafka APIs, clients, and tooling while gaining lakehouse-native storage that eliminates the operational burden of managing Kafka clusters. **Key differentiators:** - Native Kafka service — not a protocol translation layer; no code changes required - Leaderless architecture eliminates partition rebalancing and broker failures - Diskless design with lakehouse-native storage (Apache Iceberg, Delta Lake) reduces infrastructure cost by up to 95% in benchmarks vs. traditional broker-based Kafka - Direct on-ramp to data lakehouse — streams become queryable tables without ETL - Cloud-native auto-scaling with no capacity planning - Unified semantics across Kafka and Pulsar workloads on a single engine ### Pulsar Service - URL: https://streamnative.io/data-streaming/pulsar - Category: Messaging & Streaming Apache Pulsar as a fully managed service for enterprise messaging and streaming workloads. Pulsar provides native multi-tenancy, geo-replication, and a unified model for both queues and streams. **Key differentiators:** - Production-proven Apache Pulsar — adopted at enterprises like Intuit, Splunk, and Iterable - Native multi-tenancy with tenant/namespace isolation - Built-in geo-replication across regions and clouds - Unified queues and streams in a single system — no need for separate message brokers - Modernization path from legacy messaging (RabbitMQ, ActiveMQ, IBM MQ) - Lakehouse-native when running on Ursa Engine ### Universal Connect (UniConn) - URL: https://streamnative.io/data-streaming/connect - Category: Data Integration UniConn is a unified connector framework that brings together Kafka Connect and Pulsar IO under a consistent, declarative interface. It enables organizations to connect, process, debug, and monitor data pipelines across 200+ production-ready connectors. **Key differentiators:** - Single interface for both Kafka Connect and Pulsar IO connectors - 200+ pre-built connectors for databases, data warehouses, SaaS applications, and cloud services - Declarative pipeline configuration - Built-in monitoring, debugging, and error handling - Runs alongside Kafka and Pulsar workloads on the same platform ### Universal Linking - URL: https://streamnative.io/data-streaming/universal-linking - Category: Data Replication & Lakehouse Integration - Availability: Public Preview Universal Linking mirrors any Kafka-compatible cluster into Ursa with full fidelity, then automatically represents those streams as Iceberg or Delta tables in your lakehouse. It enables real-time replication and lakehouse integration without ETL pipelines. **Key differentiators:** - Mirror any Kafka-compatible cluster (Confluent, Amazon MSK, self-managed) into StreamNative - Zero-ETL lakehouse integration — topics become Iceberg/Delta tables automatically - Full-fidelity replication preserving topic structure, partitions, and offsets - No vendor lock-in — standard open formats throughout - Gradual migration path from existing Kafka infrastructure ### Orca Agent Engine - URL: https://streamnative.io/agent-engine/overview - Category: AI Agent Infrastructure - Availability: Private Preview — contact StreamNative for access Orca is the secure, governed runtime for AI agents to access and act on enterprise data. It provides event-driven agent execution, live streaming context, safe tool access, and full governance — all built on the streaming data backbone. **Key differentiators:** - Event-driven runtime — agents triggered by real-time data events, not just API calls - Live context from streaming data — agents see current state, not stale snapshots - Secure tool access with fine-grained permissions and audit trails - Enterprise governance with policy enforcement, rate limiting, and compliance controls - Native integration with Kafka and Pulsar data streams - Multi-agent orchestration and coordination ### Lakestream — The Open Stream Storage - URL: https://streamnative.io/lakestream - Category: Open Stream Storage Lakestream is the open stream storage for the Streamhouse: streams stored on object storage, with a defined integration into the lakehouse. Streamhouse and lakehouse architectures are built on the same open infrastructure. The lakehouse needed tables to be open; a Streamhouse also needs the stream to be open, and Lakestream provides that open stream storage. Ursa is StreamNative's implementation, and native Kafka and Pulsar run on it. - **Stream storage (implemented by Ursa)**: Leaderless and diskless stream storage on object storage, with a write-ahead log for low-latency streaming and data compacted to Parquet for analytics. - **Stream catalog**: Streams, schemas, and tables are governed together and registered in the catalogs your lakehouse already uses. - **Protocol serving**: Stateless Kafka and Pulsar servers read and write the same streams. **Key differentiators:** - Streams sit next to Apache Iceberg and Delta Lake tables on the same object storage, without a separate stream-to-table pipeline - Open formats throughout (Apache Iceberg, Delta Lake) - Interoperability solved in storage: the protocol is a choice of interface, not a choice of data silo - Significant cost reduction vs. traditional streaming infrastructure through elimination of broker-attached storage ### Managed Apache Flink - URL: https://streamnative.io/products/flink - Category: Stream Processing - Availability: Early Access Fully managed Apache Flink powered by the Ververica VERA engine. Deploy Flink with native Kafka and Apache Pulsar in the same VPC for end-to-end stream processing. **Key differentiators:** - Powered by Ververica VERA engine with enterprise optimizations - Auto-scaling based on workload demand - Flink SQL support for accessible stream processing - Runs in the same VPC as Kafka and Pulsar workloads - 24/7 expert support from streaming specialists - Integrated monitoring and job management ## Cloud Deployment ### StreamNative Cloud - URL: https://streamnative.io/cloud/overview StreamNative Cloud is the fully managed platform for running data streaming and AI agent workloads. It offers three deployment models to match different security, compliance, and operational requirements: - **Serverless**: Instant provisioning, pay-per-use, zero infrastructure management. Best for development, testing, and variable workloads. - **Dedicated**: Single-tenant clusters with dedicated resources, custom configurations, and SLA guarantees. Best for production workloads requiring isolation. - **BYOC (Bring Your Own Cloud)**: StreamNative control plane manages clusters running in your AWS, Azure, or Google Cloud account. Best for strict data residency and compliance requirements. **Supported clouds:** AWS, Microsoft Azure, Google Cloud Platform ## Solutions - URL: https://streamnative.io/solutions StreamNative solutions address real-time data and AI agent use cases across industries including financial services, technology, retail, logistics, gaming, and IoT. Common patterns include event-driven architectures, real-time analytics, Kafka modernization, legacy messaging migration, and agentic AI applications. ### Confluent Alternative - URL: https://streamnative.io/confluent-alternative For organizations seeking an independent, cloud-native Kafka platform, StreamNative offers native Apache Kafka on Ursa Engine with lakehouse-native storage, up to 95% lower infrastructure cost, and zero code changes required. Following IBM's acquisition of Confluent, StreamNative provides a vendor-independent alternative with a leaderless, diskless architecture that eliminates common Kafka operational challenges while providing native dual-protocol support for both Kafka and Pulsar. ## Customer Success Stories StreamNative serves enterprises across industries who rely on the platform for mission-critical real-time data infrastructure: - **Discord**: Built a real-time streaming ML platform on managed Pulsar, Flink, and Iceberg to power safety and personalization for 150M monthly active users. - **Intuit**: Built next-generation messaging platform on Apache Pulsar, replacing legacy infrastructure. - **Flipkart**: Implemented topic-as-a-service using Apache Pulsar for India's largest e-commerce platform. - **Netdata**: Enabled unlimited infrastructure monitoring powered by Apache Pulsar's streaming capabilities. - **Transport Exchange Group (TEG)**: Modernized logistics data infrastructure with Apache Pulsar and StreamNative. Full case studies available at: https://streamnative.io/success-stories ## Company ### About StreamNative - URL: https://streamnative.io/about StreamNative was founded by the original creators of Apache Pulsar at Yahoo. The team includes core committers and PMC members of Apache Pulsar, Apache BookKeeper, and other open-source projects. The company is backed by leading investors and is headquartered with a globally distributed team. ### Open Source Heritage StreamNative is deeply rooted in open source. The founding team created Apache Pulsar and continues to be the largest contributor to the Pulsar ecosystem. The company maintains a commitment to open-source standards and avoids vendor lock-in through the use of open formats (Apache Iceberg, Delta Lake) and open protocols (Kafka, Pulsar). ## Frequently Asked Questions **What is a Streamhouse?** A Streamhouse is a shared data architecture for the age of AI. Streamhouse architectures capture, transport, transform, govern, and serve the current state of a business continuously, so that production applications, analytics, and AI agents can act on it. They are real-time, production-native, and decentralized. Streamhouse and lakehouse architectures are built on the same open infrastructure — object storage, open table formats, and catalogs: the lakehouse is data to analyze the business, and the Streamhouse is data to run it. Learn more at https://streamhouse.com. **How does StreamNative deliver a Streamhouse?** StreamNative is the Streamhouse for apps, analytics, and AI agents. The StreamNative Data Platform captures data with Universal Connect and Universal Linking, transports it on native Kafka and Pulsar, transforms and serves it with SQL Workspace (Private Preview) and Apache Flink (Early Access), governs it in the catalogs you already run, and stores streams on Lakestream, the open stream storage, with Ursa as its implementation. See https://streamnative.io/data-platform/overview. **What is Ursa Engine?** Ursa is StreamNative's storage engine and its implementation of Lakestream, the open stream storage. It powers Kafka and Pulsar services and is leaderless (no partition leader elections or rebalancing), diskless (data stored in object storage and lakehouse formats), and lakehouse-native (streams are automatically represented as Iceberg/Delta tables). **Can I use my existing Kafka applications?** Yes. StreamNative provides a native Kafka service, not a protocol translation layer. Existing Kafka producers, consumers, Kafka Streams applications, and Kafka Connect connectors work without code changes. **What is Lakestream?** Lakestream is the open stream storage for the Streamhouse: streams stored on object storage, with a defined integration into the lakehouse. Because Streamhouse and lakehouse architectures stand on the same open infrastructure, streams and Apache Iceberg or Delta Lake tables share the same ground instead of living in separate systems. Ursa is StreamNative's implementation of Lakestream, and native Kafka and Pulsar run on it, so the protocol becomes a choice of interface, not a choice of data silo. **How does StreamNative differ from Confluent?** StreamNative provides native Kafka — not just compatibility — on a single lakehouse-native engine (Ursa), eliminating the need for separate systems for streaming and messaging. Following IBM's acquisition of Confluent, organizations seeking an independent alternative benefit from StreamNative's leaderless, diskless architecture with built-in Iceberg/Delta integration without ETL, dual-protocol support (Kafka and Pulsar), and up to 95% lower infrastructure cost. **What is the Orca Agent Engine?** Orca is StreamNative's runtime for AI agents. Unlike generic agent frameworks, Orca is built on the streaming data backbone, giving agents event-driven triggers and live context from real-time data streams, with enterprise-grade governance and security. **Is there a free tier?** Yes. StreamNative offers a free Serverless tier for development and evaluation at https://console.streamnative.cloud/signup.