TaubyteThe biggest hidden cost of a new platform is rarely the platform. It's the migration.
The pitch isn't "replace everything," it's "run what you already built, faster and cheaper, at your own pace."
Bring your own data layer
Every database resource is typed by what it holds: kv, document, sql, or vector, and each type picks its own engine.
Use Taubyte's own distributed engine
Replicated across the swarm, zero setup, zero servers to run.
Replicated across the swarm: kv, document and vector run this way today.
Zero setup: No servers to run yourself.
Or point at what you already run
sql connects to an external PostgreSQL or MySQL server with credentials pulled from a secret, over TLS with full certificate verification by default. document can reach an external MongoDB. The same pattern extends to messaging: stream runs on gossipsub or an external Kafka/NATS cluster.
PostgreSQL / MySQL over TLS: Credentials pulled from a secret.
MongoDB for document: Point at what you already run.
What else keeps working as-is
Your file storage
Speaks the same S3-shaped API (get/head/put/delete/list, etags, continuation tokens).
cache
A fast, replicated key-value cache with per-entry TTL and a choice of consistency.
sync
An exactly-once queue or append-only log for coordination between workloads.
What you already run
Read natively, no manual translation.
tau import
Same compiler your team uses every day.
Native Taubyte Resources
Reports exactly what didn't carry over.