Some workloads need more RAM per CPU than our standard plans provide.
Memory-optimized plans double the memory-to-CPU ratio at a lower price per GB.
Pricing · Aug 18, 2026
Meet our high-memory instance types
High-memory instance types for caches, data processing and ML inference.
Keep reading
Product · Oct 2, 2026
Why long-running agents need checkpoints
Long-running agents need checkpoints, retries and observability. Here's how we built Workflows to provide them.
Pricing · Sep 24, 2026Key Value just got cheaper
We've reduced Key Value prices across every plan, effective immediately.
Product · Sep 12, 2026One build, many services: shared artifacts for monorepos
Monorepos with many services can now share a single build artifact.
Start building on Nubra today
Join millions of developers and teams. Free plan available, no credit card required.