Rivet Actors adds SQLite storage: one database per agent, tenant, or document

Rivet Actors, an open-source alternative to Cloudflare Durable Objects, has released SQLite storage under Apache 2.0 license. Every actor gets its own SQLite database, allowing for millions of independent databases—one per agent, tenant, user, or document.
Key Details
This approach provides per-entity data isolation with SQLite running in-process with each actor. A custom VFS persists writes to HA storage (FoundationDB or Postgres). The single-writer actor model keeps reads local and fresh, unlike client-server databases where reads can be slow or stale.
Use cases include:
- AI agents: per-agent DB for message history, state, embeddings
- Multi-tenant SaaS: real per-tenant isolation without RLS hacks
- Collaborative documents: each document gets its own database with built-in multiplayer
- Per-user databases: isolated, scales horizontally, runs at the edge
How it compares
- Cloudflare Durable Objects & Agents: similar colocated SQLite and compute, but closed-source and vendor-locked
- Turso Cloud: closed-source with different use case—clients query over network, so reads are slow or stale
- D1, Turso (the DB), Litestream, rqlite, LiteFS: tools for running a single SQLite database with replication, while Rivet is for running lots of isolated databases
Rivet Actors also provide realtime (WebSockets), React integration (useActor), horizontal scalability, and actors that sleep when idle.
📖 Read the full source: HN AI Agents
👀 See Also

Cull: Open-Source Dataset Curation Engine for AI Image Pipelines
Cull scrapes images from 340+ sources including Civitai, X/Twitter, Reddit, Discord, and booru sites, classifies them with a vision-language model via local LM Studio or Groq, and sorts into category folders with SD prompts and audit records.

Using Claude Code to revive abandoned personal projects: a practical walkthrough
Matthew Brunelle shares how he used Claude Code (with Opus 4.6) to resurrect a stalled YouTube Music–to–OpenSubsonic API shim project, complete with setup steps, prompts, and workflow tips.

cortex-engine MCP server adds persistent memory and multi-agent support
cortex-engine v0.4.0 is an open-source MCP server that gives AI agents persistent long-term memory with tools like observe(), query(), believe(), and dream(). It now supports multiple agents with isolated memory namespaces.

Career-Ops Fork Adds LinkedIn Job Discovery Using Apify
A developer forked the career-ops Claude Code system and added LinkedIn job discovery using Apify, addressing the main limitation of the original project which only scanned pre-configured company career pages.