//opengauntlet
// surveyed, not benchmarked

Vector databases and AI agent memory compared

A raw vector database can find the nearest match; deciding what's worth remembering across sessions, resolving a fact that changed, and forgetting what's gone stale is a different, opinionated layer on top. Here is every vector database, agent-memory framework, and graph-memory system we could verify against a primary source, on the axes that decide a project.

Surveyed, not benchmarked. Nothing on this page was measured by OpenGauntlet's judge pipeline.

How to read this section

None of this is scored by the judge pipeline. OpenGauntlet measures Conversational Language Humanlikeness in text. This section is a sourced survey of vector databases and memory systems, not a trial. Where a claim could not be verified against a primary source, it says so rather than smoothing it over.

Scale and performance figures are mostly vendor-reported. This project does not run its own retrieval benchmarks. Where a row's scale claim could not be independently confirmed, its evidence field says so rather than presenting a vendor number as measured fact.

//

Every system, compared

Filter by kind, pick the columns you care about, search any field, click a heading to sort. Every row carries the one thing a buyer would otherwise find out too late.
Loading…
commercial-safe weights permissive, with a catch not usable commercially proprietary service discontinued
//

Claims this corrects

Each of these is repeated widely, and each is wrong.
//

What we couldn't establish

A comparison that hides its gaps is less useful than one that names them.