AI Packing List: How AI Can Decide What to Pack

Search for an AI packing list and you'll find plenty of tools that produce one. Most of them are doing something narrower than the name suggests: generating text that looks like a packing list. That's genuinely useful for not forgetting your charger. It's not the same as deciding what to pack.

Quick Answer

Generation vs Decision

Ask a language model for a packing list for five days in Lisbon in October and you'll get a good one. It knows the climate, it knows what people do there, and it can assemble a sensible set of categories: light layers, a rain shell, comfortable walking shoes, four or five tops.

That output is built entirely from general knowledge about Lisbon in October. Nothing in it is about you. It doesn't know whether your four tops are merino or heavy cotton, which changes how many you need. It doesn't know you own a rain shell already. It doesn't know your bag is 32 litres.

Deciding requires a different shape of input. You need a set of candidates, which means a catalogue of what you actually own. You need constraints: the trip's temperature range and activities, how many days, whether laundry is available, how much space you have. And you need a way to score each candidate against those constraints so that some items win and others get cut.

That's an optimisation problem rather than a text generation problem. Both can be built with machine learning, and both produce a list at the end, which is why the difference is invisible from outside.

What Deciding Actually Requires

A wardrobe catalogue. This is the input almost nothing has, and it's the reason most AI packing tools stop at categories. It's also the real cost of the approach: someone has to photograph or import their clothes once before any of this works.

Wear-count reasoning rather than day-count reasoning. How many tops you need for a week isn't seven, and it isn't a fixed four either. It depends on how many wears each garment delivers, which depends on fabric, destination temperature, and activity. Merino goes further than cotton. A system that selected specific garments can compute this; a system that output "four shirts" cannot, because it doesn't know what those shirts are.

Combination checking. Packing the correct number of items is easy. Packing items that work together is the thing that determines whether you wear what you brought. That requires knowing what each piece looks like: colour relationships, formality level, and which pairings are simply wrong regardless of score.

A capacity ceiling. A list with no volume limit will happily exceed your bag, which is why the last step of most packing apps is you, on the floor, taking things back out.

Where This Goes Wrong

The most common failure is a tool that personalises the language rather than the decision. Output that says "for your Lisbon trip, pack four versatile tops" reads as tailored while containing exactly as much information about you as a generic list.

The second failure is treating every category the same. Underwear scales one-to-one with days. Tops scale with wear count. Bottoms barely scale at all, since trousers go several wears before washing. A system applying a uniform per-day multiplier gets two of those three wrong, which is how people end up with seven pairs of trousers for a week.

The third is ignoring the constraint that actually binds. On a European budget airline the limit isn't a 40-litre carry-on, it's a 24-litre under-seat bag. A list that doesn't know which one you're working with can't be right by accident.

What JetKit Does, In a Nutshell

JetKit is built as a decision layer rather than a list generator, and it works in roughly four steps.

It starts from your closet. You catalogue what you own once, with photo recognition handling most of the tagging. You can also photograph a full outfit and have it separated into individual garments. That catalogue is the candidate set everything else operates on.

It scores candidates against the trip. Each item is evaluated across several dimensions rather than a single relevance score: how well it suits the occasion, the destination's weather, how it coheres with the rest of the selection, your own stated preferences, practicality, versatility across multiple days, and how recently you've worn it.

It applies hard rules before scoring. Some combinations are wrong regardless of how well they score, so they're eliminated outright rather than ranked. Swimwear with dress shoes, shorts with boots, formal pieces with casual footwear. This is deliberately a filter rather than a preference, because a scoring system alone will occasionally produce something absurd with a high number attached.

It sets quantities from trip structure. Trip length selects a packing formula, laundry access reduces the counts, and your chosen packing mode adjusts the whole thing toward minimal or comfortable. That's what makes the difference between a ten-day list and a seven-day list roughly one top rather than 43 percent more of everything.

The output names specific garments from your closet and confirms they assemble into complete outfits, which is the step a category list structurally can't reach.

For a comparison of how existing tools handle this, including which ones actually read your wardrobe, see our roundup of packing apps that use your wardrobe.