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

@memberjunction/context-crush

v5.47.0

Published

Dependency-light token-optimization primitives: structural JSON compression, cache-aligned prompt prefixes, and AST-aware code reduction

Readme

@memberjunction/context-crush

Dependency-light, framework-agnostic token-optimization primitives for AI agents and any TypeScript code that ships large payloads into a model's context window. Every primitive is a pure function over plain strings/objects — no MemberJunction entity coupling, no AI calls, deterministic output.

What's inside

| Export | Purpose | |---|---| | CrushJSON(value, opts?) | Structural JSON compression — collapses arrays-of-objects to a columns/rows table, elides null/empty fields, interns repeated long strings, and emits byte-stable minified output. | | DescribeCrush(result) | Renders a compact, model-readable legend explaining a CrushJSON result so the meaning is recoverable. | | PartitionStablePrefix(messages, opts?) | Splits a conversation into a byte-stable stable prefix and a mutable volatile tail, so callers can prune/compact without invalidating a provider's KV-cache prefix. | | CrushCode(source, lang, opts?) (subpath @memberjunction/context-crush/code) | AST-aware source reduction — keeps signatures/structure, collapses non-focal function bodies. Behind a subpath export so the base import stays parser-free. |

Usage

import { CrushJSON, DescribeCrush } from '@memberjunction/context-crush';

const result = CrushJSON(largeArrayOfRecords, { MaxChars: 8000 });
const prompt = `${DescribeCrush(result)}\n${result.Text}`;
// result.OriginalChars / result.CrushedChars report the savings.
// Code reduction lives behind a subpath so the base package stays zero-dep.
import { CrushCode } from '@memberjunction/context-crush/code';

Determinism

CrushJSON produces byte-identical output for the same logical input (stable key ordering, no Date.now/random). This is load-bearing: it lets PartitionStablePrefix keep a re-rendered payload from breaking provider KV-cache hits across turns.

Credits & prior art

Several primitives here are independent TypeScript re-implementations of concepts from Headroom (chopratejas/headroom), an Apache-2.0 context-compression system for AI agents:

  • CrushJSON / DescribeCrush ← Headroom's SmartCrusher
  • PartitionStablePrefix ← Headroom's CacheAligner
  • CrushCode ← Headroom's CodeCompressor

These are clean-room implementations written from the published algorithm descriptions — no Headroom source is copied, vendored, or depended upon, and this package does not use the "Headroom" name in its own identity. See the NOTICE file and the canonical attribution record in plans/agent-token-optimization.md (§0).