Replikativ

Replikativ is an umbrella for open systems focused on persistent state, structural sharing, and speculative computation.

The projects published under the org.replikativ namespace explore how databases, indexes, and coordination systems can be built as immutable, isolated verifiable structures rather than mutable, interferring systems.

The open-source foundations are maintained by Christian Weilbach, whose research in simulator-based inference and probabilistic programming (UBC PLAI Lab) informs the design of these systems. Intelligent systems need memory that versions, persists, and branches — just as probabilistic inference requires tracking distributions over time.

Projects

datahike.io — The memory layer for intelligent systems. Immutable databases, vector search, and analytics with git-like versioning.

• Datahike — an immutable Datalog database with time-travel and branching
• Proximum — version-controlled HNSW vector indexing
• Scriptum — git-like branching for Apache Lucene full-text search
• Stratum — SIMD-accelerated columnar analytics

Open source

Source code and artifacts are published under the replikativ GitHub organization and the org.replikativ Maven coordinates (was io.replikativ in the past, please upgrade).

For questions, commercial inquiries, or to get involved, visit datahike.io or contact contact@datahike.io.

© Christian Weilbach