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

@iharkharytanovich/found

v0.1.9

Published

Foundation — AI Domain Knowledge CLI (published bundle).

Readme

Foundation

A workbench for research and cross-disciplinary knowledge work — agents specialized on your curated knowledge domains, not another chatbot.

Foundation renders results in the reply — a full cheminformatic analysis of aspirin with a 2D structure, computed properties, and the domain's compute kits

Foundation renders the result, not just text: here a full cheminformatic analysis with a 2D molecule, computed properties, and the domain's offline compute kits — one of 14 built-in scientific renderers.

Most of the leverage an LLM delivers in a discipline comes not from better model weights but from the right context around it. Foundation builds, stores, and connects that context. Three primitives:

  • Domain — the crystallized essence of a discipline, structured by a Domain Context Contract (DCC): a frontmatter schema + ranked search + index hierarchy. The agent navigates hundreds of knowledge files as fluently as a coding agent navigates a million lines of code. Domains are grown by the agent and versioned through Git.
  • Workspace — the raw material (papers, datasets, notes). One workspace attaches to several domains at once.
  • Project — a graph of specialists wired by CCTP contracts (sync / async / fan-out), with transports as the I/O boundary: inbound webhooks and CLI invoke your domains, agents call external APIs and scripts.

A project as a graph of cooperating domains wired by CCTP contracts, with transport in/out nodes at the edges

A project is a graph of cooperating domains wired by CCTP contracts, with transports as the I/O boundary.

Try it in a minute

Requires Node.js 22 LTS and npm (not pnpm or Bun — they break native modules like better-sqlite3).

npm install -g @iharkharytanovich/found
found web                # open the UI at http://127.0.0.1:3737

found web starts a local daemon and opens the UI. Configure your LLM provider (Anthropic / Google / OpenAI / Ollama / OpenAI-compatible) on the Settings page — keys land in the OS keychain or a 0600 dotfile, no shell setup. For an always-on local setup that survives reboots, run found web pin once.

The bundle already ships the aristotle domain (137 structured files) — open it and ask "what does Aristotle think about friendship and virtue?"; the agent doesn't hallucinate, it navigates a real knowledge base and quotes specific passages. Shipping alongside it is the foundation domain — ask it directly what the app can do, how to set it up, and what's possible.

Per-domain Git tab — push your domain changes and pull colleagues' context; here 11 changed context files ahead of remote, ready to Push

Per-domain Git tab: push your domain edits and pull colleagues' context without leaving the UI — expertise is a portable Git artifact.

Why this isn't "a chatbot wrapper"

  • Local-first — your data and domains stay with you (Git artifacts, not our cloud).
  • Any provider — Anthropic / Google / OpenAI / Ollama / OpenAI-compatible.
  • Offline computation — Python (numpy/scipy/sympy) in a WASM sandbox.
  • Cheap long sessions — prompt caching cuts input tokens by up to 70–85% on long sessions.
  • Transparency — every run, sub-agent, and CCTP step is visible and reproducible.

Full guide

📘 Full guide & feature walkthrough → — the DCC and knowledge structure, projects/CCTP/transports, Git collaboration, scientific rendering, offline computation, cron tasks, MCP, and self-hosting.

Self-host via Docker

For server / headless deployment, Foundation runs in Docker as a thin wrapper around this same npm bundle. There is no prebuilt image — every host builds locally from three small reference files (Dockerfile, docker-compose.yml, Caddyfile). The Dockerfile does npm install -g @iharkharytanovich/found@${FOUND_VERSION} at build time, runs found serve as a non-root user under tini, and Caddy fronts it on two listeners: :80 for the full UI + API (LAN / VPN only) and :8080 for the public webhook receiver (/api/in/* paths, Standard Webhooks 1.0 HMAC-SHA256 authenticated).

The fastest way to obtain the three reference files: install the CLI, then ask the bundled foundation domain agent for them.

npm install -g @iharkharytanovich/found
found web                        # open the UI, pick the `foundation` domain
                                 # → ask: "give me the reference Dockerfile,
                                 #    docker-compose.yml and Caddyfile"

Save the three files into a fresh directory, then:

docker compose build             # one-time, ~1–3 min (installs from npmjs.com)
docker compose up -d

Open http://foundation.localhost in your browser. *.localhost is reserved by RFC 6761 for loopback — no /etc/hosts or DNS setup required.

To pin a specific version instead of latest:

FOUND_VERSION=0.1.1 docker compose build --no-cache
docker compose up -d

State persists in a named Docker volume (found-data/data). Use a named volume, not a bind mount — SQLite WAL locking does not work reliably over Docker Desktop's macOS / Windows file-sharing layer.

To stop: docker compose down. To upgrade: docker compose build --pull && docker compose up -d.

Update

npm install -g @iharkharytanovich/found@latest

The CLI prints a one-line notification when a newer version is available.

Release Notes

Uninstall

npm uninstall -g @iharkharytanovich/found
rm -rf ~/.found        # Remove local data (conversations, settings, domains)

If you ran found web pin, run found web unpin first to remove the OS supervisor registration.

License

@iharkharytanovich/found is licensed under the PolyForm Noncommercial License 1.0.0 — free for personal, academic, research, and educational use.

Commercial use (SaaS hosting, paid services, embedding into commercial products, or internal use at a for-profit organization) requires a separate commercial license. Contact: [email protected]

See the LICENSE and NOTICE files for the full license text and copyright / trademark information.