Four products, each with its limits named
All four are in production. Each one states where it stops, because a named limit is cheaper to discover here than in your evaluation.
Vyrell
LiveConnectors authenticate as low-privilege, read-only audit readers and poll each engine’s native audit trail across ten engines, scoring every event against 215 rules.
You get the audit trail and the compliance report without installing anything on the production database.
The Sybase ASE and Cassandra connectors are built but not yet supported on Kubernetes deployments.
Nothi
LiveA catalog-driven control plane discovers the database operators installed on your cluster and provisions each database as a custom resource: PostgreSQL, MySQL, MariaDB, MongoDB, Valkey, Kafka, RabbitMQ, OpenSearch, Oracle, EDB, MySQL Community, and Qdrant.
One console for twelve engines, with namespace-per-tenant isolation, per-engine backups, hourly metering, and no per-engine platform code to add the next engine.
A thirteenth engine, openGauss, is built but not offered. On storage classes that cannot expand, a storage grow is refused up front rather than half-applied.
Dorbar
LiveEvery room is exactly one of two classes, fixed when it is scheduled. In an Official room the in-cluster media server is the trust anchor, so recording, post-meeting transcription, offline live captions, and host moderation are available. A Sealed room is end-to-end encrypted, MLS (RFC 9420) group keying over a post-quantum X-Wing suite driving SFrame media encryption, so the server only ever sees ciphertext.
A Sealed room refuses every server-side feature that would touch content, and refuses to join rather than fall back to plaintext. An Official room is protected in transit and is never described as end-to-end encrypted.
Sealed-room encryption runs at the browser tier: post-quantum confidentiality on stock browser crypto, medium assurance pending external cryptographic review, with a FIPS-140-3 native client a later phase. Recording cannot be paused mid-session, and local recording is deliberately not built.
Moncho
LiveManaged Kubernetes on infrastructure we operate: a hosted control plane per tenant, workers joined from a pre-staged pool, and three isolation tiers, up to a cluster that shares no API server, etcd, or kernel with any other tenant.
A bank can take the cluster to its vendor-risk board: posture rolled up per framework, sealed into dated tamper-evident snapshots, with continuous CIS scanning and OSCAL export.
Adding a node joins a pre-staged machine, not elastic autoscale. No financially-backed 99.9% SLA until a second-site DR drill is published. Restricted-data workloads wait on the in-country substrate.
Live means the product is complete and running for customers. Where a capability inside a live product is planned rather than shipped, it appears under that product’s Limits and nowhere else. No claim on this site that the code does not support.