How to Use Claude as a Travel Budget Planner: The Elena Skill Guide

Updated
Skill · .md

The skill behind this guide: Elena - Travel Budget Planner AI Skill. Holds the real prices you looked up, what you actually spent last trip, and how you travel, so each budget is built from your own numbers rather than the internet's average - $19, one payment, yours permanently.

View the Elena skill →

Trip budgets almost never fail on the estimate. They fail because the cheap flight landed at one in the morning, and the airport taxi and the extra night were not in the flight column. A budget is not a list of prices, it is a chain of decisions, and most of the money you did not plan for is the second-order cost of a choice you already counted as a saving. A model cannot tell you what anything costs. It is unusually good at tracing what a decision commits you to.

0:15 · loop Same AI, same question. One file makes it a specialist.

The costs you missed were consequences, not surprises

Look at where a holiday actually overruns and the pattern is consistent. It is rarely a single shock. It is a set of small, entirely predictable costs that were created by earlier decisions and then filed under the wrong heading, or under none.

The flight that saved you sixty euros arrives after the last train, so it costs a taxi. The hotel that saved you thirty a night is twenty minutes out, so it costs two journeys a day for six days, and it costs you eating near the hotel rather than where you wanted to eat. The early check-in you did not pay for means a day with luggage, which means the luggage locker and a longer lunch. None of these are surprises. They were all determined the moment you booked.

So the useful question is not "what will things cost" but "what has each decision already committed me to". That question a model can genuinely help with, because it is reasoning about consequences rather than recalling prices, and reasoning about consequences is a thing it does well.

Never let it give you the first number

Here is the part that matters more than anything else on this page, and it is not the usual warning about accuracy.

A model has no live prices. What it has is an average absorbed at training time, across seasons, neighbourhoods, traveller types and a currency that has since moved. So the figure is wrong. That is the small problem.

The large problem is that the figure becomes your reference point. Once it has told you to budget eighty a day for food, you have an anchor. Every real price you then see gets judged against a number that was invented, and you will make actual trade-offs against it - skipping the restaurant you came for because it feels like an overspend, or feeling comfortable for five days and broke on the sixth. An inaccurate estimate you can correct. An anchor rearranges your decisions without you noticing.

The fix is not "check its numbers". It is do not let it produce a number at all until you have supplied some. Look up three real prices yourself - one flight, one night's accommodation, one ordinary meal - and make everything else a proportion of those. You are then using the model for what it is good at, which is knowing that a trip has a laundry line and an airport-transfer line and a Sunday-closing problem, and not for what it cannot do, which is knowing what any of them cost this month.

Prompt 1 - the decision audit

Run this before any figure exists. It is the one that finds the money.

Do not give me any prices or estimates. Not one number. Here are the decisions I have already made about this trip: [FLIGHTS WITH TIMES, ACCOMMODATION AND WHERE IT IS, HOW LONG, WHO IS GOING, WHAT I HAVE BOOKED] For each decision, tell me what it commits me to spending later that I have probably not counted: - Arrival and departure times: what is open, what is running, and what does the gap force? An early arrival before check-in and a late departure after check-out both cost something. - Location of the accommodation: how many journeys a day does this create, and does it change where we will end up eating? - What the accommodation does NOT include: breakfast, kettle, laundry, towels at the beach, a fridge - Any cheap option I chose: what is the second-order cost of that saving? - The days with nothing planned: those are the expensive ones, because unplanned time gets filled by whatever is nearest Then: which single decision I have already made will cost me the most in consequences I have not budgeted for? Say it plainly even if it is too late to change it, because I would rather budget for it than be surprised by it.

Prompt 2 - the lines, with no figures in them

List every cost line this specific trip will have. No amounts. A blank spreadsheet, not a budget. Group them: - Before we go: visas, vaccinations, insurance, gear, parking at the home airport, pet or house cover - Getting there and back: including every transfer at both ends - Sleeping: including tourist taxes and deposits - Eating: and split the kinds - proper meals, coffees, water, the supermarket run - Moving around locally: and say whether this place is a walk-everywhere city or not - Doing things: entry fees, booking fees, guides, equipment hire - Money itself: card fees, cash withdrawals, the small change that never gets used - Things that only apply to THIS destination and season For each line, tell me whether it is fixed, per day, per person, or per journey. That structure matters more than the amounts, because it is what changes when the trip changes. Then flag the lines people most often leave out entirely for a trip of this shape, and say why they get missed.
Skill spotlight
Elena - Travel Budget Planner AI Skill
Elena - Travel Budget Planner AI Skill
$19 · one payment

Works from anchors you supply rather than figures it invents, and keeps the part that makes budgets get better: what you actually spent last time, per line, against what you planned. After two trips you are estimating from your own history instead of from an average of strangers, which is the only version of this that ever gets accurate.

View Elena - Travel Budget Planner AI Skill →

Prompt 3 - build from three real prices

Go and look up three things first. A specific flight, one night at a place you would actually stay, one ordinary meal. Ten minutes. Then this prompt has something true to stand on.

Here are three real prices I looked up myself today: - Flight, [ROUTE, DATES]: [PRICE] - One night, [THE KIND OF PLACE I WOULD STAY, AREA]: [PRICE] - One ordinary meal out, not touristy: [PRICE] Build the rest of the budget as RATIOS to these, not as absolute figures you have decided on. Rules: - Every line must be expressed relative to one of my three anchors, and you must show the reasoning. "Coffee is roughly a fifth of an ordinary meal here" is an answer. "Coffee is four euros" is not. - Where a line cannot honestly be derived from my anchors - entry fees, transfers, the tourist tax - say so and write [LOOK THIS UP: exactly what to search for], rather than estimating it. - State every assumption you are making about how I travel, and mark the ones that would change the total most if wrong. Then give me the arithmetic in full, so I can change one anchor and see the whole thing move. I want a model I can argue with, not a total.

Prompt 4 - a buffer that is an answer, not a percentage

Adding fifteen per cent is a ritual. A buffer only works if you know what it is for.

Do not tell me to add a percentage buffer. For this specific trip, name the three most likely ways it costs more than planned. Real ones, from the shape of this trip - not generic advice. For each: - What happens - Roughly what it costs, expressed against my anchors - Whether it is avoidable now, at booking, or only manageable once it happens - What I would actually do in the moment Then tell me: what single amount would let me absorb the most likely two without changing the trip? That is my buffer, and I want it derived rather than rounded. Finally, the uncomfortable question: is there a version of this trip that fits my real budget without a buffer at all - one fewer night, a different month, a nearer destination? Answer it honestly even though I have clearly already decided.

The four per cent nobody budgets: how you pay

Card and cash handling is a real line on a trip and it never appears in a budget article. Three things to settle before you go.

Always choose the local currency at the terminal. When a card machine abroad offers to charge you in your own currency instead, that is dynamic currency conversion, and the rate is set by the merchant's provider rather than by your bank. Which? documented a Ryanair booking where €156.45 was converted at 93p to the euro while Visa's rate that day was 87p - about £9 on one transaction. Their advice is blunt: "Always pay in euros or any other local currency... It's better to let your card provider or bank make the currency conversion". Source: Which?

Know your own card's foreign-transaction fee before you go, from your bank rather than from a model, and know whether cash withdrawals are charged differently from purchases. And check whether your destination is card-first or cash-first, because arriving with the wrong assumption costs you either ATM fees all week or a wallet of currency you convert back at a loss.

Prompt 5 - the money-handling layer

Separately from the trip budget, help me plan how I will pay for things in [DESTINATION]. Tell me what to check, and name the source for each. Do not state fees, rates or limits yourself - you do not know my bank and you do not know today's rates. Cover: - Is this place card-first or cash-first in practice, and where specifically will I still need cash? - What do I need to ask my own bank about: foreign transaction fees, ATM withdrawal charges, whether debit and credit differ - Whether to take cash out at home or there, and what decides it - Tipping: whether it is expected, roughly how it works, and whether it is cash even where payment is card - Any local payment method a visitor will hit - transport cards, apps, machines that take only coins Then write me a short pre-departure checklist of things to do with my bank and my cards, in the order to do them, with enough lead time flagged where something takes days.

Prompt 6 - reconcile, so the next one is yours

The only prompt here that makes the next trip more accurate rather than this one. It is also the honest answer to why a persisted skill beats a fresh chat for this particular job.

I am back. Here is what I actually spent, by line: [PASTE] Here is what I planned: [PASTE] For each line: - Planned against actual, and the ratio between them - Was the gap a bad estimate, or a decision I made on the trip? Those are different problems with different fixes. - Which lines were I consistently wrong about, and in which direction? Then: - Update my anchors: what a meal, a night and local transport actually cost me, in that place, in that month - Which of the forgotten-cost lines did I hit, and which never materialised for this style of travelling? - What would I change about how I budget rather than about how I travel? Keep this as a record with the destination and month attached, so the next trip starts from what I spent rather than from an average of strangers.

Where it fails

Failure What happens What to do about it
Anchoring An invented daily figure becomes the number you judge every real price against Supply three real prices first. Never let it produce the first number
Stale prices An average across seasons and neighbourhoods, from a currency that has since moved Ratios to your anchors, and a look-this-up line for anything that cannot be derived
Currency conversion It converts at a rate it half-remembers and presents the result as a figure Do currency conversion yourself, on the day, with a live rate
Category thinking A tidy budget by heading that misses everything created by a booking decision Audit the decisions first. Most missed money is a consequence, not a category
Ritual buffers Add 15% contingency Name the three likely overruns for this trip and size the buffer to absorb two
Optimism about you It budgets for the traveller you describe, not the one who buys a round every evening Give it last trip's actual spending. It cannot know you and will not guess pessimistically
Fees are invisible Card fees, withdrawals and conversion never appear as a line at all Plan the paying separately from the spending, with your own bank as the source
No memory A fresh chat cannot know you always underestimate food and overestimate transport Reconcile after every trip and keep the record where the next session can read it
One-off alternative
Travel Budget Calculator Prompt
Travel Budget Calculator Prompt
$9 · one payment

One trip, booked, leaving soon. A copy-paste prompt that runs the line structure and the anchor arithmetic in a single pass. No record kept between trips, which is exactly the trade you are making at this price.

View Travel Budget Calculator Prompt →

How do you install the Elena skill?

The download is a ZIP with SKILL.md at the root of the archive, not inside a nested folder - that folder structure is the usual reason an upload fails. In the Claude desktop app, open Customize → Skills, upload the ZIP and toggle it on.

Skills need code execution enabled, under Settings → Capabilities. Anthropic's help centre currently lists Skills on Free, Pro, Max, Team and Enterprise, while its Academy tutorial lists Pro, Max, Team and Enterprise - so if you are on the free plan, check Settings → Capabilities for your own account rather than taking either page's word for it. The full walkthrough is in the skill installation guide.

In ChatGPT or Gemini there is no upload step: open SKILL.md, copy the contents, and paste them into custom instructions. You lose automatic triggering and keep the method.

Who is this for?

People deciding whether a trip is affordable before they commit rather than after, anyone travelling on a number that cannot move, and households where two people have different ideas about what the holiday costs. It works in Claude, ChatGPT or any AI chat.

The trip splits into its own skills, each $19 and a one-time download:

The wider set is the travel, food & home skills.

In summary:

Budgets fail on consequences rather than estimates, so audit the decisions you have already made before you write down a single figure - the cheap flight that lands after the last train has a taxi and a night attached to it. Never let the model produce the first number, because an invented figure does not just mislead you, it becomes the anchor you judge real prices against and quietly rearranges your choices. Look up three real prices yourself and make everything a ratio to those, with a look-this-up line for anything that cannot be derived. Replace the ritual percentage buffer with the three ways this particular trip is likely to overrun. Plan how you will pay separately from what you will spend, and always choose the local currency at the terminal. Then reconcile when you get home, so the next budget is built from what you actually spent. For anchors and spending history held between trips, Elena - Travel Budget Planner AI Skill ($19); for one trip leaving soon, the Travel Budget Calculator Prompt ($9). Both work in Claude, ChatGPT and any AI chat, with a 30-day money-back guarantee.

Skill · .md · Works with Claude & ChatGPT

Elena - Travel Budget Planner AI Skill

Audits what your bookings already commit you to, builds from anchors you supply instead of figures it invents, sizes the buffer to the trip rather than to a percentage, and keeps what you actually spent so the next estimate is yours. No subscription. Yours permanently.

$19
Get the Elena skill →

KissMySkills is a marketplace of 2,000+ AI skills, 300+ prompt packs, 100+ agents & free tools for Claude, ChatGPT & any AI chat.

Related Skill Guides

Put this into practice

Or get every skill with Library See plans →

Frequently asked questions

Can Claude plan a travel budget?+

It can do the structural half well and the numeric half not at all. It knows a trip has an airport-transfer line and a laundry line and a Sunday-closing problem, which is genuine recall over a huge body of trip write-ups. It does not know what any of those cost this month, because it has no live prices - only an average absorbed at training time, across seasons, neighbourhoods and traveller types, in a currency that has since moved. Use it for the lines and the consequences, and supply the numbers yourself.

Why do trip budgets always go over?+

Rarely because of a bad estimate. Almost always because of second-order costs created by a decision you already counted as a saving. The flight that saved sixty euros lands after the last train, so it costs a taxi. The hotel that saved thirty a night is twenty minutes out, so it costs two journeys a day for six days and changes where you end up eating. None of these are surprises - they were determined at booking. The useful question is not what will things cost but what has each decision already committed me to.

Why should I not let the AI give me a price first?+

Because the damage is not the inaccuracy, it is the anchor. Once it tells you to budget eighty a day for food, every real price you see gets judged against a number that was invented, and you start making real trade-offs against it - skipping the restaurant you came for because it feels like an overspend, or feeling comfortable for five days and broke on the sixth. An inaccurate estimate you can correct. An anchor rearranges your decisions without you noticing.

What is the best way to get a realistic number?+

Look up three real prices yourself before you prompt anything: one specific flight, one night at the kind of place you would actually stay, one ordinary non-touristy meal. Ten minutes of work. Then have the model build everything else as ratios to those three anchors, showing its reasoning, with an explicit look-this-up line for anything that cannot honestly be derived - entry fees, transfers, tourist taxes. You get a model you can argue with rather than a total you have to trust.

How big should a travel budget buffer be?+

A percentage is a ritual rather than a plan. Fifteen per cent that one missed connection would swallow is not a buffer. Ask instead for the three most likely ways this particular trip overruns, what each costs relative to your anchors, whether it is avoidable at booking or only manageable once it happens, and what you would actually do in the moment. Then size the buffer to absorb the two most likely without changing the trip. Derived rather than rounded.

Should I pay in my own currency or the local one abroad?+

The local one, always. When a card machine offers to charge you in your home currency instead, that is dynamic currency conversion and the rate is set by the merchant's provider rather than by your bank. Which? documented a Ryanair booking converted at 93p to the euro when Visa's rate that day was 87p, about nine pounds on a single transaction, and their advice is to always pay in the local currency and let your card provider do the conversion.

What money costs do budgets always miss?+

The handling itself. Foreign transaction fees on your own card, whether cash withdrawals are charged differently from purchases, ATM fees, currency you convert back at a loss, and whether the destination is card-first or cash-first in practice. None of this appears in a normal budget and a model will not raise it. Plan how you will pay separately from what you will spend, and get the fee answers from your own bank rather than from a chat window, because it does not know your bank.

Is a skill better than a one-off prompt for budgeting?+

Only if you travel more than once or twice a year, and then clearly yes, for one reason: reconciliation. A budget only becomes accurate when it is built from what you actually spent last time rather than from an average of strangers. That requires a record of planned against actual, per line, with the destination and month attached. A one-off prompt produces a decent structure for the trip in front of you and remembers none of it.

~/get-started

Skills that work. No fluff.

Browse every skill, prompt pack, and agent in the store.

Browse all skills →Or try the free tools