VeloDB Cloud (Apache Doris)

VeloDB is a company built by the original creators of Apache Doris, the open-source real-time analytical database, and operates internationally through VeloDB Technology International Pte Ltd, an IMDA-accredited company in Singapore. It serves organizations that need sub-second analytics, log search and AI retrieval at scale without running several separate engines.

VeloDB offers VeloDB Cloud as a fully managed SaaS or bring-your-own-cloud service on AWS, Azure and GCP, VeloDB Enterprise for self-managed deployment on VMs or Kubernetes, and the open-source Apache Doris. One SQL engine queries structured, JSON, full-text and vector data, with separated storage and compute, columnar compression and materialized views. It reads lakehouse tables in Iceberg, Delta Lake, Paimon and Hudi through catalogs such as Unity, Polaris and AWS Glue. Users include Xiaomi, ByteDance, Kwai, MiniMax and Ave.ai.

• Real-time analytics: Sub-second queries under high concurrency, heavy updates and complex joins.
• Hybrid search: Vector, full-text and structured queries combined in one SQL statement.
• Lakehouse access: Queries Iceberg, Delta Lake, Paimon and Hudi via Unity, Polaris, Glue or Hive Metastore.
• Observability stack: Log storage and search with OpenTelemetry and Grafana as an ELK alternative.
• Deployment choice: SaaS, BYOC on AWS, Azure or GCP, or self-managed Enterprise edition.
• Ecosystem connectors: Spark, Flink, Kafka, dbt, BI tools and AI frameworks such as LangChain and Dify.

A Vietnamese bank or telecom operator can replace an ELK stack for log analytics and security monitoring at lower storage cost. An e-commerce or fintech app can run user-facing reports and search on the same engine. A company building AI assistants can store embeddings next to business data for retrieval.

Analytics, full-text search and vector search run in one SQL engine, cutting the number of systems to operate. The vendor reports observability migrations from ELK reducing cost by 80% and an AI company storing about 83% less than with Elasticsearch. Pricing is hourly by vCPU from $0.134 for SaaS and $0.11 for BYOC, with a 30-day trial and $300 in credits.

Apache Doris has a large user base in Asia, and enterprises are consolidating analytics, log search and vector retrieval to control cost as AI applications grow. Open lakehouse formats and the wish to avoid lock-in to proprietary cloud warehouses are also driving adoption of open-source analytical engines.

Built by the original creators of Apache Doris. Listed in the IMDA directory of innovative tech companies in Singapore as an IMDA-accredited company; the Singapore entity is VeloDB Technology International Pte Ltd. The Apache Doris community roughly doubled in 2025 according to the company.