About ShannonBase
ShannonBase is an open source database from Shannon Data AI that combines transactional and analytical work in one system. It is wire compatible with MySQL 8.4, so existing drivers, ORMs and dump files work unchanged. It keeps a columnar mirror of InnoDB tables in memory and routes each query to whichever engine is faster, which the vendor describes as analytics without ETL or a separate warehouse.
Beyond analytics, it adds AI features inside the database. These include a native vector type with an index and embeddings generated in process, training and prediction from plain SQL using LightGBM and ONNX Runtime, local LLM generation with retrieval augmented generation, and an agent runtime in which writes proceed only after human approval. The site shows a read only playground and a published TPC H benchmark of a developer build, with a reproducible harness. The current release is labeled version 0.1, and a Docker image is available.
It is aimed at engineering teams running MySQL based applications who want analytics, vector search and machine learning on the same data. It is released under GPL v2.
The software is free to use. The vendor's pricing page shows a generic Personal, Startup and Enterprise table that does not describe the database, and the vendor publishes no ShannonBase specific prices.