Stuff Stash

Caution

Stuff Stash is under early and active development.

By using it, you acknowledge that you’re on the bleeding edge, and agree to not hold me accountable for any cuts you may obtain 🩸.

I built Stuff Stash because I wanted an inventory system but didn’t have time to maintain it.

While Stuff Stash can be maintained by clicking, tapping, and typing, I use the LLM-powered voice mode because I value my time.

I made this for me and my home. Stuff Stash might not be a good fit for you. That’s fine. You might consider HomeBox, Snipe-IT, or even Grocy.

Stuff Stash supports cloud and local models for text to speech, speech to text, and LLMs.

Screenshots

…If you insist.

The design is subject to change (read: it absolutely will change, for the better.)

But why not share where I’m at:

1 / 1

Slide 1 of 1

Why It Exists

My previous inventory solution (shoutout HomeBox ❤️) was invaluable when I needed to find something I had in storage.

But when I needed to move things around, add new items, or lend something to a friend, it fell short.

It wasn’t for lack of functionality, but rather the friction was too high.

Core Tenets

• Low-friction upkeep: inventory lifecycle must be manageable with native voice control.
• Mobile App: the inventory must be manageable via a cross-platform native mobile app from day one.
• Local control: everything (even “AI” stuff) must work offline and be self-hostable.
• Native SSO (OIDC) Support: OIDC support must be natively integrated from the start.
• Multi-tenancy and sharing: a multi-tenant, inventory, and user model must be natively supported.
• Portable data: data must be easily imported and exported.

Start Here

If you want to see how to run Stuff Stash, start with Run Stuff Stash.

If you want to learn more about the internals, see What It Does.

Note

A note on the use of AI

I imagine hope we’ll look back on this time in history and laugh one day.

But as of 2026, the use of AI in software development is still quite polarizing, and developers are often praised or hanged for their transparency about the use of AI.

Here are a few things I want you to know:

• I value your time, so I spent my time writing this entire page (how can I prove that? I can use any emoji I want 🙈 🍘 🪘 ⛓️ 🪥 🆗 🅱️0️⃣0️⃣Ⓜ️𝓮®️)
• Most of the documentation is generated at this time. This is likely to change as the product matures and I shift my effort towards the new user experience. For now, you’ll have to deal with AI docs. At least there are docs :)
• Stuff Stash’s code was 100% generated by an LLM (primarily gpt5.5 via Codex), driven by spec-driven development.
• Good security and architecture design are a primary concern and are where I spend most of my mental effort. For example:
  • I enforce a minimum age for packages and libraries via a commit hook.
  • All package, library, base image, and GitHub Action use pinned versions.
  • All container images are signed and come with build provenance attestations.
• Without the use of “AI” coding tools, this project would never have come to be. I think it’s better that this project exists. You have the right to disagree, but I find this useful.