npm package discovery and stats viewer.

Discover Tips

  • General search

    [free text search, go nuts!]

  • Package details

    pkg:[package-name]

  • User packages

    @[username]

Sponsor

Optimize Toolset

I’ve always been into building performant and accessible sites, but lately I’ve been taking it extremely seriously. So much so that I’ve been building a tool to help me optimize and monitor the sites that I build to make sure that I’m making an attempt to offer the best experience to those who visit them. If you’re into performant, accessible and SEO friendly sites, you might like it too! You can check it out at Optimize Toolset.

About

Hi, 👋, I’m Ryan Hefner  and I built this site for me, and you! The goal of this site was to provide an easy way for me to check the stats on my npm packages, both for prioritizing issues and updates, and to give me a little kick in the pants to keep up on stuff.

As I was building it, I realized that I was actually using the tool to build the tool, and figured I might as well put this out there and hopefully others will find it to be a fast and useful way to search and browse npm packages as I have.

If you’re interested in other things I’m working on, follow me on Twitter or check out the open source projects I’ve been publishing on GitHub.

I am also working on a Twitter bot for this site to tweet the most popular, newest, random packages from npm. Please follow that account now and it will start sending out packages soon–ish.

Open Software & Tools

This site wouldn’t be possible without the immense generosity and tireless efforts from the people who make contributions to the world and share their work via open source initiatives. Thank you 🙏

© 2026 – Pkg Stats / Ryan Hefner

@blueshed/delta

v0.4.11

Published

Delta-doc — JSON-Patch document sync with SQLite and Postgres backends, WebSocket transport, and reactive client signals

Readme

@blueshed/delta

Use delta-doc for shared state. That sentence is the whole pitch.

Three op verbs (add / replace / remove), one path grammar (/coll/id/field), one WebSocket. Three backends ranked by complexity (JSON file → SQLite → Postgres). The browser code never changes when you move up the ladder.

It's smaller than the alternative you're reaching for. The whole system fits in one context window — your AI can read all of it before writing a line.

The recipe — one command, three files, two browser tabs

bun examples/shared-state/server.ts

A live multi-user chat. Two browser tabs at http://localhost:3100. Type in one, it appears in the other. Persists to a JSON file. No database, no schema, no codegen, no fetch calls.

The whole sync layer:

// server.ts
import index from "./index.html";
import { createWs, registerDoc } from "@blueshed/delta/server";

const ws = createWs();
await registerDoc(ws, "chat:room", { file: "./chat-room.json", empty: { messages: {} } });

const server = Bun.serve({
  port: 3100,
  routes: { "/": index, [ws.path]: ws.upgrade },
  websocket: ws.websocket,
});
ws.setServer(server);
// client.ts
import { connectWs, openDoc } from "@blueshed/delta/client";
import { applyOpsToCollection } from "@blueshed/delta/dom-ops";

const doc = openDoc("chat:room", connectWs("/ws"));

await doc.ready;
// ... initial paint from doc.data.get() ...
doc.onOps((ops) => applyOpsToCollection(log, "messages", ops, { create, update }));

// Send: one verb, one path.
await doc.send([{ op: "add", path: `/messages/${crypto.randomUUID()}`, value: { author, text } }]);

Walk into examples/shared-state/ for the complete, runnable version.

Choose a backend (only when you need to)

Same client (openDoc / doc.send / doc.onOps), same op verbs across all three. Default to JSON-file; graduate only when forced.

| Tier | Pick when | Server | |---|---|---| | JSON file | Single doc, single process, prototyping. | registerDoc(ws, name, { file, empty }) from @blueshed/delta/server | | SQLite | Many docs, relational queries, temporal history. | registerDocs(ws, db, schema, docs, customDocs?) from @blueshed/delta/sqlite | | Postgres | Cross-process fan-out, RLS, stored-function auth. | createDocListener(ws, pool, { custom? }) + registerDocType(...) from @blueshed/delta/postgres |

Browser code does not change when you graduate.

Custom doc types — predicate-based views

defineCustomDoc(prefix, opts) declares a read-only doc whose contents are decided by a Bun-side predicate over a watched collection. Useful for bbox queries, tag filters, anything you'd express as WHERE in a live materialised view. SQLite and Postgres only. See examples/sites-bbox/ for both backends.

CLI

After install, the delta bin is invokable via bunx delta:

# Talk to a running server
bunx delta open  <docName>             # one-shot open + print + exit
bunx delta watch <docName>             # stream broadcast ops
bunx delta delta <docName> <opsJSON>   # apply ops
bunx delta call  <method>  [paramsJSON]

# Postgres setup
bunx delta init init_db --with-auth
bunx delta sql ./types.ts --out init_db/003-tables.sql

URL resolution: --urlDELTA_WS_URL.delta file in cwd → ws://localhost:${PORT:-3100}/ws.

Layout

core.ts       — JSON-Patch applyOps + types (zero deps)
client.ts     — reactive doc + reconnecting WS (peerDep: @blueshed/railroad)
server.ts     — WS action router + JSON-file backend (registerDoc, registerMethod)
sqlite.ts     — SQLite backend (schema codegen, temporal tables, doc lenses, custom docs)
logger.ts     — tagged, level-gated console logger
postgres/     — schema · sql · listener · registry · stored-function SQL
examples/
  shared-state/ — the canonical "use delta-doc for shared state" recipe
  sites-bbox/   — custom doc types, both backends
  kanban/       — full Postgres + auth example
  todos-vs-rls/ — RLS / auth-jwt walkthrough

Exports

| Subpath | Runs | Purpose | |---|---|---| | @blueshed/delta/core | anywhere | applyOps, DeltaOp | | @blueshed/delta/client | browser | connectWs, openDoc, call, WS | | @blueshed/delta/dom-ops | browser | applyOpsToCollection — keyed-DOM op routing | | @blueshed/delta/server | Bun | createWs, registerDoc (JSON-file backend), registerMethod | | @blueshed/delta/sqlite | Bun | defineSchema, defineDoc, defineCustomDoc, registerDocs(..., customDocs?), snapshots | | @blueshed/delta/postgres | Bun + pg | defineSchema, defineDoc, defineCustomDoc, generateSql, applyFramework, createDocListener(ws, pool, { custom? }), registerDocType, docTypeFromDef, withAppAuth | | @blueshed/delta/logger | anywhere | createLogger, setLogLevel, loggedRequest | | @blueshed/delta/auth | Bun | DeltaAuth interface, wireAuth, upgradeWithAuth | | @blueshed/delta/auth-jwt | Bun + pg + jose | jwtAuth({ pool, secret }) — reference impl |

Vendor-first

Delta's framework SQL is copied into your init_db/ by bunx delta init, not read from node_modules at runtime. The philosophy is shadcn/ui for database schemas:

  • Files are explicit, git-tracked, and yours to read.
  • Upgrades run delta init --upgrade (with .bak backups) and tell you what changed.
  • Your setup.ts walks init_db/ in alphabetical order — no hidden imports.

TypeScript runtime is imported normally. Only SQL is vendored.

Why

Existing sync libraries are built for human developers: big API surfaces, many idioms, ecosystem dependencies. Delta-doc is shaped for AI-driven development — the whole system fits in context, has one way to do each thing, and generates its schema from a single TypeScript source of truth.

If your AI is reaching for Supabase or Firebase, that's a default trained from millions of projects. Delta-doc isn't harder than those — it's smaller. Your AI can read all of it.

Lineage: started as dzql (Vue / Pinia, database-first), matured into seiro (CQRS over WS with Preact Signals), refined in paintbrush's delta-sync, realised in clean as a Postgres-resident primitive. This package is the extraction.

For Claude

The skill at .claude/skills/delta-doc/SKILL.md is the runbook. It triggers on "shared state", "real-time", "live updates", "multi-user", "collaborative", and on names of competing libraries. Install the package and Claude Code discovers it automatically.

Status

Published as @blueshed/delta on npm — see package.json for the current version and CHANGELOG.md for what's in each release. The shape is stable; new backends and custom doc types are additive.