Built on ProximaDB

AnvaiOps is open-core: the retrieval engine is open source, and the commercial layer is the managed control plane, governance, and economics around it.

The two halves

	ProximaDB (open source)	AnvaiOps (commercial)
What	The context database — hybrid vector + BM25 + graph in one engine	Control plane + policy layer over it
Provides	Storage, indexing, retrieval, query	Accounts, connectors, tenant isolation, metering, MCP, console
Runs	Anywhere you run it	Managed in your cloud
License	Open source, auditable	Commercial

Where the line is

The engine measures and retrieves — units in, results out. AnvaiOps adds the things a managed multi-tenant service needs and the OSS engine deliberately leaves out:

• Entitlement — who can see and query what (tenant isolation).
• Governance — data-quality grades, sensitivity classification.
• Economics — usage metering and cost attribution.
• Integrations — the connector fleet that fills the engine with your knowledge.

flowchart TB
    subgraph OSS["ProximaDB — open source"]
        E[Engine: vector + BM25 + graph + SQL]
    end
    subgraph COM["AnvaiOps — commercial layer"]
        G[Entitlement · governance · metering · MCP]
        C[Connectors · console · control plane]
    end
    C --> G --> E

Why open-core

You can inspect exactly what stores and retrieves your data — the engine is auditable and self-hostable. What you pay for is the managed, governed, metered service around it, not a black box.

The engine's capability maturity is the ceiling for what AnvaiOps claims: a feature never advertises a higher status than the engine actually supports.