Next Steps

You’ve got a running todo app. From here, follow one of these threads depending on what you’re trying to add.

Understand what plasma is doing

• Mental model — the three layers plasma is built on
• Schema and mutators — how the isomorphic types actually work
• Optimistic vs canonical — why the client sees mutations before the server does, and what reconciles them
• Push, pull, rebase — the sync loop in detail
• clientID vs clientGroupID — the identity model
• Change log and cookies — the data structures the sync loop hands around

Grow the schema

• Schema DSL — every column type, modifier, and option
• Refs and cascades — foreign keys, onDelete, subscript ergonomics
• Migrations — how ensureSchema and runMigrations handle schema evolution

Handle multi-user data

• Auth and permissions — auth.read, auth.write, and how they run on both sides
• Conflict resolution — when the server row and the client row can’t be reconciled by rebase alone

Add files and rich state

• Files and blobs — the file() column type, R2 storage, and usePlasmaFile
• CRDT columns — G-Counter / PN-Counter / LWW-Register / OR-Set
• Encryption — .encrypted() columns, DEK wiring, at-rest vs E2EE trade-off

Ship it

• Deployment — Wrangler, D1, R2, environment wiring
• Testing — fake-indexeddb for the client, Miniflare
  • @cloudflare/vitest-pool-workers for the server
• Devtools — the in-page panel and the Chrome DevTools bridge

Reach for advanced features

• Offline mode — PlasmaClientOptions.offline and per-table changeLogSuppressed
• Multi-region — the SequencerDO and the causal cookie
• Live queries in detail — subscribeDelta, whenReady, IVM eligibility

Real apps

• Todo app — the Quick Start built out into a complete deployable project
• Offline notes with files — offline mode + image attachments + R2

Something broke

• Troubleshooting — indexed by symptom. If you got an error message, start here.