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

@superdesign/loopbase

v0.2.0

Published

Cross-session agent memory — discover, read, search, and hand off work across local AI agent sessions (Claude Code, Codex, pi).

Readme

loopbase

Cross-session agent memory. Discover, read, and hand off work across local AI

agent sessions — Claude Code, Codex, and pi — on one machine.

agent-facing usage in SKILL.md.

What it does

Every agent dumps its session transcript to disk. loopbase indexes all of them

(lazily, no daemon) and gives one CLI to:

  • **loopbase list** — see the sessions in your project, across agents, each with its

    worklog as a table of contents.

  • **loopbase search "<text>"** — find sessions/turns by content across agents; each

    hit carries a handle to jump straight to it.

  • **loopbase show <id>** — understand a session: a structural map by default

    (worklog + turn outline), then drill into a turn, a worklog span, a tool

    result, or a subagent.

  • **loopbase log "<what I did>"** — leave a worklog entry on your current session;

    it auto-spans the messages since your last log.

  • **loopbase cost** — token + USD cost per session (also --summary by model,

    or loopbase cost <id> for one session's per-model breakdown).

  • **loopbase insights** — ranked automation candidates mined from every

    session's tool calls: repeated/expensive patterns, recurring call sequences, and tools that keep failing. Ranked by real attributed USD, tagged with the dominant repo, and nested with sub-clusters (composio run → the Intercom tools it's really running). Each row links to show --turn examples.

  • **loopbase serve** — a local web dashboard over everything above (see below).

Output is JSON by default (--text for humans). Session id is auto-resolved from

the environment — agents pass nothing.

Web dashboard (loopbase serve)

loopbase serve              # → http://localhost:4178
loopbase serve --port 8080  # pick another port

A local, read-only dashboard over your indexed sessions — no account, nothing leaves your machine, and it re-indexes on each load so it's always fresh. What you get:

  • Sessions by cost — every session with its token + USD spend; sortable, filterable by agent and time window, with resizable columns.
  • Group by working dir — fold sessions under their project (git root), each with a cost subtotal; the git branch is shown per session.
  • Drill into a session — per-model cost breakdown with a token-burn sparkline, cost per log batch (when you've logged more than once), the worklog, and a paged conversation viewer loaded on demand.
  • Insights tab — automation candidates across all sessions: the most repeated/expensive tool patterns, recurring call sequences, and recurring errors, each with example sessions to open.

Cost is a list-price estimate computed from a built-in price catalog (subscription billing differs). Refresh rates from upstream any time with loopbase cost --refresh.

Install

loopbase runs on Bun ≥ 1.3 (it uses bun:sqlite).

bun add -g @superdesign/loopbase   # installs `loopbase` (+ `lb` alias) globally
bunx @superdesign/loopbase list    # or run without installing

Prefer a standalone binary (no runtime needed)? See INSTALL.md.

Develop

bun install
bun run loopbase -- list          # run the CLI from source
bun run serve:dev           # dashboard with hot reload (edits in src/web/ live-reload)
bun run check               # lint + typecheck + tests
bun run build               # compile a single binary → dist/loopbase

Requires Bun ≥ 1.3.

Status

v0 (the JSONL family): Claude (read + write), Codex (read), pi (read). SQLite

(hermes) and openclaw adapters, hooks, LLM summaries, and team sync are planned

License

Mostly AGPLv3; files marked /* @license Enterprise */ fall under the

loopbase Enterprise Commercial License. See LICENSE.