A directory of 74 assistants across 16 categories — who makes them, where to find them, and which ones expose an API you can actually build against.
Before you rely on this. This list is compiled from public sources, and the AI landscape moves faster than any directory. Products get renamed, merged and retired constantly — verify anything load-bearing directly with the provider. Listing here is not an endorsement, and does not imply the product integrates with Caura.
On the user numbers. A figure appears only where the vendor stated it publicly or it comes from a regulatory filing, and each one is labelled with who said it and when. Where no such figure exists the field is simply absent — we don’t republish third-party traffic estimates, which for some products on this page span an order of magnitude (Claude’s circulating estimates run from 19M to 245M, Perplexity’s from 45M to 230M). Note also that weekly and monthly actives are different metrics and are not comparable across rows.
Scroll the list and a pattern shows up: almost all of them are excellent within a single conversation, and almost none of them remember anything useful across sessions, tools, or teammates. Close the tab and the context is gone. Open a different assistant and you start over.
That gap is what Caura exists to close — a governed memory layer that sits underneath agents rather than inside any one of them, so what one agent learns is available to the next, with scopes and an audit trail attached. If you are wiring several of the tools above into one workflow, that shared layer is the piece you end up needing.
What is AI agent memory? · Persistent memory for AI agents · Get started with Caura