Launch HN: Rivet (YC W23) – Scale and Orchestrate Agents on S3
Why CEREBRO kept it
Agent orchestration infrastructure; agentic SaaS-adjacent platform launch.
The text below is an automated extraction of the article at https://github.com/rivet-dev/rivet, stored verbatim in the public cerebro-vault repository. Copyright remains with the original publisher (github.com).
4.99 ms p99 cold starts, 72 KB RAM per Actor, scales to billions Quickstart • Documentation • Blog • Discord • X Rivet Actors are long-lived processes with in-memory state, a SQLite database, WebSockets, and scheduling. You can create one for each agent, user, session, or tenant. - 72 KB per Actor: Actors share a node process instead of each booting a sandbox, so one server holds hundreds of thousands. - 4.99 ms p99 cold starts: No image pull and no VM boot. - Built for scale: All data lives in S3, so a cluster scales to billions of Actors. - Hibernates when idle, wakes on demand: Idle Actors
Community take
Commenters find the Rivet UX confusing even for coding agents, and one wants self-hostable TypeScript actors without the agent layer since Node has no proper Akka-style framework.
Backlinks
Appeared in 1 briefing