23 Sept 2026
When the model you build on stops answering
Our own count: two of the 122 AI models we publish record a date on which they stop being served. So we read nine providers’ own…
A Postgres-compatible engine that puts full-text, vector and hybrid search in the same database as columnar analytics; the single-node build is Apache 2.0 and the first production release, Krummelanke, shipped on 22 September 2026.
SereneDB is a real-time search analytics database from a Berlin company of the same name, founded in 2025 by Alexander Malandin, Andrey Abramov and Valery Mironov. Its first production-ready release, named Krummelanke, was published on 22 September 2026. The pitch is consolidation: full-text, vector and hybrid search alongside SQL analytics over columnar storage, behind a PostgreSQL-compatible front end, so a team does not run a search cluster beside an analytical warehouse and keep the two in step. The search core is IResearch, a C++ information-retrieval library that members of the team have worked on since 2016, and the company says it outperformed two established search libraries on one of their own public benchmarks — that is the vendor’s own reading of a third-party benchmark, and it is not recorded as a score here. The complete single-node version is under the Apache 2.0 licence and published on a public code host; the company states that the single-node build always will be. What the announcement does not publish is a price list for anything beyond that, so a team sizing a cluster has to ask. For a buyer the interesting question is not the benchmark but the operational one: a single engine means one backup story, one access-control model and one upgrade path, and it also means one thing to be wrong about.
An earned signal from verification, reviews, awards, transparency and engagement — the vendor can't buy it.
Updated 9/23/2026
No questions yet. Be the first to ask about SereneDB.
Everything here links back to the same verified catalogue. Pick your next stop.