General FAQs and troubleshooting
Find answers to common questions about KDB.AI. For more specific guidance, refer to the Server Setup FAQ or Performance and Optimization FAQ.
General
Q: What is KDB.AI?
A: KDB.AI is a high-performance vector database designed for efficient data storage, retrieval, and analysis.
Q: How do I get started with KDB.AI?
A: The quickstart guide provides step-by-step instructions for setting up and using KDB.AI.
Q: What are the system requirements for KDB.AI?
A: Refer to Hardware requirements and System requirements on the setup page.
Q: What vector indexes are supported?
A: KDB.AI Server supports indexes like Flat, qFlat, IVF, IVFPQ, HNSW, and qHNSW. See about indexes and how to use indexes.
Q: What data types are supported?
A: KDB.AI supports the data types listed here.
Q: Can I migrate data from another platform to KDB.AI?
A: Yes. Contact support@kdb.ai or the Slack community for migration guidance.
Q: Can an external kdb+ database be used with KDB.AI?
A: KDB.AI supports referencing data held in an external kdb+ database rather than storing the full payload internally. Common reasons to do this include:
- Prevents data duplication — your records stay in their existing store and only the vectors and a reference are held in KDB.AI.
- Access to KDB-X analytics — KDB.AI is purpose-built for similarity search and does not support the full range of analytics functions available in KDB-X. If you need fast time-series or columnar analytics alongside vector search, using KDB-X as the external database lets KDB-X handle analytics while KDB.AI handles the vector layer.
Note that using an external database does not remove the need to ingest vectors into KDB.AI when new data is added — the index still has to be kept in sync with the source.
Q: How do I contact KX Support?
A: Email support@kdb.ai or join the Slack community.
Licensing
Q: What are the restrictions of the KDB-X community license?
The community (kdb-x) license is free but subject to usage restrictions. The most common one to be aware of is the 24 CPU core limit. If your machine has more than 24 cores, add --cpuset-cpus to the docker run command to stay within the limit, for example --cpuset-cpus 1-24. Commercial licenses do not have this restriction.
Refer to the full kdb-x community edition usage restrictions for details. For commercial licensing enquiries, contact support@kdb.ai.
Troubleshooting
For setup troubleshooting (Docker, licensing, volumes, connectivity), refer to the Server Setup FAQ.
For performance issues, refer to the Performance and Optimization FAQ.
For other issues, contact support@kdb.ai or the Slack community.