A collection of any size — coins, stamps, vintage cameras, vinyl, comics, watches, antique tools, art books — eventually crosses a line. You used to know where everything was. Now there are too many items, too many receipts, too many "I should look that up later" notes about which mark on the bottom of which thing means it's the rare variant. And the spreadsheet you started a year ago has columns you don't remember why you added.

The work of collecting splits roughly into three streams that all need to live somewhere: an inventory of what you own, the provenance and paperwork for each piece, and the ongoing market research about what things are worth and what's worth chasing next. Most collectors keep these in three different places, which is why the system breaks.

A vault that holds all three, with an agent that can search across them, fixes most of the friction. (If you've never thought of your collection records as a long-term archive, [Turn Casual Captures Into a Searchable Life Archive](/guides/personal-life/casual-captures-searchable-life-archive/) makes the same case from the other direction.)

## One inventory, embedded right where you write about it

The temptation with an inventory is to build a dedicated app for it — a database with pictures, fields, categories, custom views. That's fine until you want to write a paragraph about *why* you bought the 1962 example and *how* the dealer described it. Then you need a second tool for the prose, and the prose ends up disconnected from the row.

In Docapybara, an inline database lives directly inside a page using a `:::database:::` directive. So your *Inventory* page can have a paragraph at the top describing the collection's scope, the database in the middle with one row per item, and ongoing notes below about acquisitions you're considering. Six column types — text, status, date, number, single-select, multi-select — cover everything from grade or condition to acquisition price and current estimated value.

The agent can update the database for you. *"Add the new Hasselblad I bought yesterday: serial 5G-1024801, condition EX+, paid $850, source KEH."* You describe the row, it appears. When tax season comes around and you need a list of every acquisition this year over $500, you ask the agent and the filtered list comes back.

## Provenance documents that are actually searchable

Provenance is the part that separates a hobby from a serious collection. Receipts, certificates of authenticity, auction records, prior-owner documentation, condition reports from grading services — every piece tells a story, and the story is what holds value when you eventually sell or insure.

The problem is that provenance documents are mostly PDFs. A grading certificate, a Christie's catalog scan, a notarized letter from the prior owner. PDFs are the worst format for staying organized in, because most note apps treat them as opaque attachments — you can store the file, but you can't search inside it.

Docapybara converts uploaded PDFs to markdown automatically. Drop a 12-page auction catalog onto a page and the agent can actually read what's inside it. From then on, you can ask things like *"What did the Christie's catalog say about the engraving on this watch?"* and the agent pulls the relevant paragraph and tells you which page it came from.

This changes how you can structure provenance. Each major item gets its own page, the PDFs live attached to that page, and the agent can answer questions about the documentation without you having to open every file and search by hand.

## Market research that compounds instead of getting lost

Every collector keeps a running mental list of *what's worth what* and *what to watch for*. The auction realized last month for the rare variant. The forum thread where someone explained why a particular maker's mark dates a piece to a specific decade. The dealer's price list from the last show.

Most of this lives in browser bookmarks, screenshots, and a memory that gets less reliable each year. A *Market research* page in your vault solves this if you treat it like a working document instead of a dumping ground.

Drop articles in by URL, paste in price lists, save screenshots from auction houses. The agent has a `web_search` tool that can pull live web pages with source URLs — so when you're trying to remember what the last few comparable sales went for, you can ask: *"What did first-state copies of this print sell for in the last six months?"* and the agent searches, summarizes, and cites the auction records.

Over time, this page becomes the thing you actually reference when deciding whether to buy at a show. Not a graveyard of bookmarks; a working file that gets smarter the more you feed it.

## Per-item pages where everything for one piece lives together

For pieces that matter — high-value items, anchors of a sub-collection, things you might one day write about or sell — give each one its own page. Pages in Docapybara nest inside each other with no depth limit, OneNote-style, so your *Inventory* page can have a child page for each major piece, and those pages can hold:

- A short narrative — when you got it, why it matters, what makes this example special.
- The provenance PDFs, attached and searchable.
- A few photos.
- Any restoration or repair history.
- A small inline database tracking valuation history if you have it appraised periodically.

The agent treats this nested structure as one searchable pile. So a question like *"Which of my pieces have been professionally appraised in the last three years, and what did the appraiser estimate?"* gets answered by the agent reading across all the per-item pages and summarizing.

## Insurance, taxes, and the dreaded "for the family" file

Two grim but necessary jobs every serious collector eventually faces: keeping insurance documentation current, and making sure that if something happens to you, your family can figure out what you have and roughly what it's worth.

Both jobs benefit from the same thing: a single, organized vault. The insurance company wants a list with valuations and provenance. The "for the family" file wants the same list, plus context — *what should be sold versus kept, who the trusted dealers are, which pieces have sentimental over monetary value, who to call*. (For the broader "what happens after me" file, see [Estate Planning and Will Preparation: A Calm Place to Keep It All](/guides/personal-life/estate-planning-will-preparation/).)

Build this once as a page called *For my executor* (or whatever feels right) and let the agent maintain the inventory summary section. Once a year, ask the agent: *"Generate an updated summary of my collection by category, with current estimated values, and flag any pieces I haven't had appraised in over three years."* You get a one-page summary that's good enough for an insurance adjuster, an accountant, or a family member who's never collected anything in their life.

## Quick capture for shows, estate sales, and the items that get away

Some of the best collecting happens at moments where you can't sit down and properly catalogue anything. You're at a show, a dealer has an interesting piece, and you want to remember the maker's mark and the dealer's name even if you don't buy.

A new page in Docapybara is one click and you're typing. Or you tap record, hold up the phone to capture the dealer's verbal description, and walk away with a transcript. The page exists, it has a timestamp, you can sort it later. *"Saw a really good early Heuer at booth 47. Maker had three more in the back. Card says Marcus Antiques, Tampa."*

Two months later, when you're trying to remember which dealer at which show had that piece, you ask the agent and the note comes back. Captured thoughts you used to lose to the chaos of a show floor stay searchable. The same logic shows up under [Notes for Hobbyists: From Woodworking to Astronomy in One Vault](/guides/personal-life/ai-notes-hobbyists-woodworking-astronomy/) if you're collecting and crafting.

## A starter setup that works on day one

If you're moving from a spreadsheet (or three spreadsheets) to a single vault, this is the shape we'd suggest starting with:

- **Inventory** — one page with an inline database. One row per item, with the columns that matter to your specific collection.
- **Provenance** — one parent page with child pages per major item. PDFs attached. Searchable.
- **Market research** — one page that grows over time. Articles, price lists, screenshots, agent-pulled web summaries.
- **Insurance & estate** — one page with an annually-refreshed summary, generated by the agent.

That's four pages. No tag taxonomy, no template requirements. The agent does the searching across them when you have a question that crosses categories.

The thing collecting eventually teaches you is that the artifacts outlast the memory of how you got them. A vault that holds the things, the paper trail, and the thinking you did along the way is what turns a collection into something that holds together long term.

[Try Docapybara free](/accounts/signup/) — start with one inventory database and add the rest as the collection grows.