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

ensemble-review

v1.0.56

Published

Dashboard for tracking your team's GitHub PR activity across repos. Runs on your machine with your gh login.

Readme

Ensemble (ensemble-review)

A local dashboard for tracking your team's GitHub PR activity across multiple repositories. Run it on your machine, see at a glance who's doing what, which PRs are stalled, which review threads are sitting untouched, and what needs your attention right now — all rendered as an interactive timeline.

There's no hosted backend, no database server, and no stored credentials. It authenticates by shelling out to your already-logged-in gh CLI, syncs activity into a local SQLite file, and serves the dashboard from a single local process.

100% free. Every feature, no paid tier, no sign-up (see License for the one limit, on large organisations).

Renamed. This package used to be limn-review (and before that pierre-review). The command is now ensemble; both old packages only forward to this one, and say so. On first start, an existing ~/.limn (or ~/.pierre-review) is moved to ~/.ensemble — your database, settings and clones come with it. If ~/.ensemble already exists, nothing is moved. Settings named LIMN_… still work; their new names start ENSEMBLE_….

Quick start

npx ensemble-review

This starts the server, prints a local URL, and opens your browser to it.

Or install globally and use the short command:

npm install -g ensemble-review
ensemble

Review and fix with Claude — free, on your own Claude

Claude Review (a code review of any PR, follow-ups, a check against a user story you paste, and a chat about the review) and AI Fix are free. By default they use your local Claude Code login: if you are signed in to Claude Code, there is nothing to set up. An Anthropic API key is the alternative: set ANTHROPIC_API_KEY. Ensemble stores no key and charges nothing for it. They can run on Codex instead (below).

  • Claude and Codex work out of the box. The first time Ensemble starts, it downloads both in the background (Claude about 110 MB, Codex about 360 MB unpacked) into ~/.ensemble/ai-runtime, at the exact versions this release was tested with. The package itself stays small. Signing in is the only step. ensemble ai install codex runs the same download by hand; ENSEMBLE_AI_AUTO_INSTALL=false skips the automatic one.
  • Set ENSEMBLE_CLAUDE_PATH to the claude you already have to use it instead of the bundled copy (the download then skips it).
  • The reviewer reads code. It cannot edit files, run commands or reach the web.
  • The fixer edits files. It has no shell, and builds and tests nothing.
  • Nothing is posted or pushed until you press the button.
  • Use my API keys first (Settings → AI), or ENSEMBLE_AI_USE_API_KEY=true, uses ANTHROPIC_API_KEY even when you are signed in to Claude Code.
  • ENSEMBLE_AI_DISABLED=true turns every AI feature off.

Codex instead of Claude

Every AI feature can run on Codex, on your own ChatGPT plan. One provider runs everything at a time.

Codex comes with Ensemble. Sign in from Settings → AI, or:

ensemble ai login codex     # sign in with ChatGPT (add --device-auth on a headless machine)

Ensemble keeps its own Codex sign-in in ~/.ensemble/codex; your ~/.codex is not read or changed. Codex with an OpenAI API key is not available yet.

  • Pick the provider in Settings → AI, or set ENSEMBLE_AI_PROVIDER to claude, codex or auto (the default: Claude when both are signed in, Codex when only Codex is).
  • Each run is labelled with the provider and model that ran it.
  • ensemble ai status shows whether Codex is set up and signed in.
  • ensemble ai logout codex signs Ensemble's Codex out.

Prerequisites

  • Node.js ≥ 20 (with npm).

  • GitHub CLI, installed and authenticated. Install it from https://cli.github.com, then run:

    gh auth login

    ensemble reads your team's activity using your gh token. It pre-checks this on startup and exits with a friendly message if gh is missing or not authed.

Native module note: ensemble-review depends on better-sqlite3, a native addon. npm installs a prebuilt binary for common platforms; if none matches your Node/OS/arch, npm compiles it from source on install (needs a C++ toolchain — Xcode Command Line Tools on macOS, build-essential + Python on Linux, MSVC build tools on Windows).

Usage

ensemble [options]
ensemble status [options]
ensemble ai install

| Flag | Env | Default | Description | |------|-----|---------|-------------| | --no-open | NO_OPEN | — | Don't open the browser on start. | | --port <n> | PORT | 4000 | Port to listen on. | | --db <path> | DATABASE_URL | ~/.ensemble/pierre-review.sqlite | SQLite DB path. | | — | ENSEMBLE_DATA_DIR | ~/.ensemble | Where Ensemble keeps its data. | | — | PERSIST_BODIES | false | Store full comment/PR text in the DB instead of loading it on demand (larger DB, but PR detail works fully offline). | | -h, --help | — | — | Show usage. |

Examples:

ensemble --port 4123 --no-open
ensemble --db /tmp/ensemble.sqlite

Data directory

Everything Ensemble keeps lives in ~/.ensemble:

~/.ensemble/pierre-review.sqlite   the database
~/.ensemble/clones/                git clones Claude Review and AI Fix read
~/.ensemble/ai-runtime/            the AI runtime, once you set AI up

The directory is created automatically. Override it with ENSEMBLE_DATA_DIR, or just the database with --db / DATABASE_URL. No team activity data or credentials are ever sent anywhere — everything stays on your machine.

How it works

Once running, open the printed URL (default http://localhost:4000). Add the repositories you want to watch from the in-app picker; the app syncs their PR activity (full backfill on first sync, incremental every few minutes thereafter) into your local DB and renders it as a timeline.

Repos are grouped into Workspaces, and the selected workspace is the scope for everything — the feed, the timeline, the metrics, the review-bot settings. Every repo belongs to exactly one workspace; new ones land in Default (renameable, not deletable), and you can create more and move repos between them from Manage repos & workspaces.

To keep the database small and backfills fast, bulky text — PR/comment/review bodies and diff hunks — isn't stored; it's fetched from GitHub (using your gh token) the first time you open a PR and cached in your browser, so re-opening an unchanged PR is instant and offline. Set PERSIST_BODIES=true to store that text locally instead — a larger database, but PR detail then works fully offline.

License

Source-available under the Ensemble License 1.0 — see LICENSE. Ensemble is free for individuals, and for organisations where no more than 3 people use it (affiliates count together). More users than that need a licence — ask through the contact form. You may not offer Ensemble, or a product built from it, as a competing commercial product or service. The AI runtime Ensemble downloads on first use is Anthropic's, under Anthropic's own terms; it is not part of this package.