Skip to main content
Glama

Compare cities

compare_cities
Read-onlyIdempotent

Rank any number of cities for a person or household: a verdict, the score out of 100 from their weights with points and reasons per line, must-pass lines as pass, fail or unknown, each earner's take-home pay and what is left after rent and spending, who can move there, how the set was scored, and what to add next. Every figure carries its source and period. Compact by default; detail "full" adds every tax line, childcare, health, leave and break-even.

Tool call template

Replace each placeholder with a value matching the input schema. Only required parameters are included; add optional parameters as needed.

{
  "id": 1,
  "jsonrpc": "2.0",
  "method": "tools/call",
  "params": {
    "arguments": {
      "cities": "<cities>"
    },
    "name": "compare_cities"
  }
}

Input Schema

NameRequiredDescriptionDefault
onlyNoScore only these criteria; every other line gets weight 0 (each listed one keeps its `weights` value, else 100). For a one-line question: only ["nice_days"] with comfort {min:20,max:28}
workNoThe person's kind of work, only if they said it: software (Software and IT), engineering (Engineering, not software), science (Research and science), medicine (Doctors and dentists), nursing (Nursing and care), teaching (Teaching and education), finance (Finance and accounting), law (Law), marketing (Marketing, design and media), office (Office and administration), sales (Sales and retail), hospitality (Hospitality and food), trades (Construction and trades), transport (Transport and logistics), public (Public sector), management (Management), founder (Founding a company). Adds what people in that work earn there and how many jobs its field holds to Work and ambition, and shares the group's weight by the kind's recipe (founder: young companies, developers, graduates); each city then says why in plain sentences (work). Leave out for the same Work and ambition as everyone
limitNoHow many ranked cities to describe in the summary (default 10); the ranking counts the rest
citiesYesCity ids (e.g. turin-it) or names in any common spelling (Turin, Torino, 'Paris, FR'); a name two places share needs its country
detailNo"summary" (default): verdict, ranking, weighted lines with citations, money in brief, who can move there, notes, required unknowns and next inputs. "full": every line, tax line and source, for programs
presetNoStarting weights
salaryNoYour gross yearly salary, the same everywhere
comfortNoWeather: what a nice day is, for criterion nice_days: the daytime temperature range, °C (default 18-30 on the feels-like temperature; basis "air" uses the air in the shade), then the tests a day must pass. Only the person's own words: rain and sky are on unless they drop them; humidity, wind and daylight only when they ask. All from ERA5 (Open-Meteo, CC BY 4.0) and the latitude
housingNo
missingNoIn fixed scoring, missing lines do not count for that city. Only with opt-in relative scoring: lowest scores gaps at the lowest compared score, worst at 0, skip drops the line for all. Money is never imputed
pensionNoShare of your gross paid into a pension, 0.05 = 5%: it leaves take-home pay and is shown as saving, not cost. Tax relief is modelled for CA (RRSP), CH (pillar 3a), CZ (pension savings (penzijní spoření)), DE (occupational pension (Entgeltumwandlung)), EE (third pillar), ES (plan de pensiones), FR (PER (plan d'épargne retraite)), GB (workplace or personal pension), HU (voluntary pension fund), IE (occupational pension or PRSA), IT (fondo pensione), JP (iDeCo), NL (workplace pension), NO (IPS (individual pension savings)), PT (PPR), SG (SRS), US (401(k)); elsewhere it is not modelled and the share is counted as saving with no relief (each city's money says which)
regimesNoNewcomer tax regime per country, ISO alpha-2 -> regime id from list_metrics_and_sources. None counts unless named here; a named one is a scenario until the person's facts and confirmations meet every condition
savingsNoSavings in the display currency, for runway
stancesNoWhere the person stands on the position lines, which score a place by its distance from them: politics 0 (left) to 10 (right); abortion 0 not allowed or only to save her life, 1 also for her health or after rape, 2 also on social or economic grounds, 3 on request in the first weeks only, 4 on request to 10 weeks or later; couples_law 0 criminalised (same-sex relations a crime), 1 legal with no union in law, 2 civil union or partnership, 3 marriage. Pass only what the person told you or what you know they hold; never guess. A stance counts once its line has a weight (they are 0 by default)
weightsNoWeight 0-100 per criterion, on top of the preset. Criteria: money_left, living_cost, right_to_live, nice_days, winter_light, sunshine, dry_days, clean_air, safety, long_life, health_costs, childcare, family_leave, happiness, builders, field_pay, field_jobs, young_companies, time_zone, flight_home, airport, public_transport, rail, short_commute, cycling, walkable, active_commute, way_out, sea, mountains, green, internet, english, international, doctors, jobs, degrees, developers, qol, clean_government, politics, abortion, couples_law, founders_run, coworking, climbing, swimming, beach, run_ride, muggy, grey, hot_nights, heat, snow, wind
childrenNoChildren's ages: childcare after public support for under-6s (OECD), child credits and allowances where the tax uses them, health for each person
personalNoYour own 1-5 lines, scores keyed by city
spendingNoMonthly spending on everything except housing, childcare and health, which are counted separately
householdNoEveryone moving, you first. Anyone after you can bring their own pay; a spouse counts for joint filing even with none
normaliseNoDefault absolute: fixed city-local scores. Relative is opt-in: adding cities can change other scores, gap handling and the shared money basis.
overridesNoYour own numbers: city -> metric id -> value; money in the city's currency. health_premium is the household's premium a month (prefer healthPremium, which takes the period); childcare_cost is the household's own childcare a month (prefer childcareCost); childcare_net_couple_1 and similar replace the OECD reference figure, a year
taxTweaksNoPer-country tax tweaks keyed by ISO alpha-2
dealBreakersNoMust-pass criterion ids, e.g. right_to_live, money_left: a city fails, passes, or is unknown (no figure, no passport or other right given); unknown ranks after every city that passes
salaryByCityNoGross yearly salary per city (offers or estimates), keyed by city id or name
childcareCostNoWhat the household actually pays for childcare in a city, after support, keyed by city id or name, in the city's currency, with per "month" or "year": replaces the OECD reference figure
healthPremiumNoThe household's health insurance premium bought outside pay (US, CH, NL), keyed by city id or name, in the city's currency, with per "month" or "year" (a year is counted as a twelfth a month)
paidChildcareNoWhether the household pays for childcare for its children under 6 (default true). false: no childcare cost is counted. Ask when it matters: the default prices full-time care for an OECD reference household
displayCurrencyNoCurrency to show amounts in (default: the salary's, else EUR)

Output Schema

NameRequiredDescriptionDefault
notesYesHow this set was scored: a line on another basis or left out for every city, periods that differ
citiesYes
detailYes
rankingYes
verdictYesWho leads, qualified when a must-pass is unknown, when the lead is provisional, and when money was compared on another basis
moneyBasisYesThe money line asked for and the basis it was compared on for every city
assumptionsYes
provisionalYesTrue when missing figures, partial money or money compared on another basis could still change who leads: say so
reportVersionYesThe report's shape version (now 2.0)
displayCurrencyYes
requiredUnknownsYes
provisionalReasonsYes

Schema Changelog

Changes observed during successful MCP inspections.

  1. Changed4 schema fields changed
    • changedInput schema / properties / comfort / description
      Previous value: -"What a nice day is, for the Nice days line: the daytime temperature range, °C (default 18-30 on the feels-like temperature; basis \"air\" uses the air in the shade), then the tests a day must pass. Only the person's own words: rain and sky are on unless they drop them; humidity, wind and daylight only when they ask. All from ERA5 (Open-Meteo, CC BY 4.0) and the latitude"New value: +"Weather: what a nice day is, for criterion nice_days: the daytime temperature range, °C (default 18-30 on the feels-like temperature; basis \"air\" uses the air in the shade), then the tests a day must pass. Only the person's own words: rain and sky are on unless they drop them; humidity, wind and daylight only when they ask. All from ERA5 (Open-Meteo, CC BY 4.0) and the latitude"
    • addedInput schema / properties / household / items / properties / entryRoutes
      Added value: +{
      +  "additionalProperties": {
      +    "type": "string"
      +  },
      +  "description": "Country to route id, only if conditions are confirmed: GB gb-ancestry (Commonwealth citizenship, eligible grandparent, support and work), gb-skilled (sponsored eligible job, salary and English); US us-h1b (cap-subject employer petition). Rough guide, not legal advice; other routes are not checked.",
      +  "propertyNames": {
      +    "type": "string"
      +  },
      +  "type": "object"
      +}
    • changedInput schema / properties / missing / description
      Previous value: -"A city without a figure on a line: \"lowest\" (default) scores it as the lowest compared city, \"worst\" as 0, \"skip\" leaves the line out for every city. Money is never filled in either way"New value: +"In fixed scoring, missing lines do not count for that city. Only with opt-in relative scoring: lowest scores gaps at the lowest compared score, worst at 0, skip drops the line for all. Money is never imputed"
    • addedInput schema / properties / normalise / description
      Added value: +"Default absolute: fixed city-local scores. Relative is opt-in: adding cities can change other scores, gap handling and the shared money basis."
  2. First observed

TDQS

A3.7/5.0
Behavior4/5

Does the description disclose side effects, auth requirements, rate limits, or destructive behavior?

The annotations already establish that this is a read-only, idempotent, non-destructive operation, so the description does not need to explain safety. It adds useful behavioral context: every figure carries a source and period, compact output is the default, and detail 'full' expands tax lines, childcare, health, leave, and break-even.

Agents need to know what a tool does to the world before calling it. Descriptions should go beyond structured annotations to explain consequences.

Conciseness4/5

Is the description appropriately sized, front-loaded, and free of redundancy?

The description is front-loaded with the core action and then enumerates output behavior efficiently, with no obvious filler sentences. The first sentence is dense and list-like, but the density is largely justified by the tool's complexity.

Shorter descriptions cost fewer tokens and are easier for agents to parse. Every sentence should earn its place.

Completeness4/5

Given the tool's complexity, does the description cover enough for an agent to succeed on first attempt?

For a 28-parameter, nested tool with an output schema and rich annotations, the description gives enough high-level context about ranking, outputs, sourcing, and compact versus full detail. Its main gap is the absence of sibling-routing guidance, but the structured metadata covers the parameter and return-value details.

Complex tools with many parameters or behaviors need more documentation. Simple tools need less. This dimension scales expectations accordingly.

Parameters3/5

Does the description clarify parameter syntax, constraints, interactions, or defaults beyond what the schema provides?

Schema description coverage is 96%, so the schema already documents nearly all 28 parameters, which sets the baseline at 3. The description adds some meaning around weighted scoring and the detail mode, but it does not meaningfully clarify most parameters such as comfort, housing, pension, regimes, or overrides beyond what the schema provides.

Input schemas describe structure but not intent. Descriptions should explain non-obvious parameter relationships and valid value ranges.

Purpose5/5

Does the description clearly state what the tool does and how it differs from similar tools?

The opening verb 'Rank' with the resource 'cities' and scope 'for a person or household' is specific and immediately distinguishes a comparison/ranking tool from lookup siblings. The following clause enumerates the verdict, weighted score, must-pass results, take-home pay, mobility, and scoring explanation, so an agent can tell what the tool produces without opening the schema.

Agents choose between tools based on descriptions. A clear purpose with a specific verb and resource helps agents select the right tool.

Usage Guidelines2/5

Does the description explain when to use this tool, when not to, or what alternatives exist?

The description states that the tool ranks cities, but it never says when to choose it over search_cities, suggest_cities, get_city, or net_pay, nor does it state exclusions or prerequisites. The implied use case is multi-city comparison, but no explicit routing guidance is given.

Agents often have multiple tools that could apply. Explicit usage guidance like "use X instead of Y when Z" prevents misuse.

Try in Browser

Glama MCP Gateway

Add one secure layer between your agents and this server.