The details

What you are wearing and carrying, what an on-foot grade or modification costs, and where its materials come from.

“what suit am I wearing” “what does night vision cost” “what would grade 5 on this suit take” “where do I find graphene”

Nothing here touches the network.

On foot is not the ship feature with a different vocabulary

Two axes, and neither one is the ship model.

Three things follow, and they run through every answer on this page.

Nothing is estimated. There is no roll count, so no hedging: every quantity below is what the game charges, exactly.

Planning matters more here than for ships. Four slots, permanent, no way to practise. A wrong modification is unrecoverable except by buying and re-upgrading a fresh item, so d47 says so once per answer rather than never.

Ordering is a routing fact. A grade 1 item has zero modification slots, and an engineer’s base has no Pioneer Supplies — only an Apex desk. So “upgrade first” is a step in the order, not a footnote: the upgrade cannot be done when you get there.

What you are wearing

get_on_foot_loadout

{"type":"object","properties":{},"required":[],"additionalProperties":false}

Wearing: Maverick Suit, grade 5 — loadout "Sneaky-Snipy".
  Suit: Improved jump assist, Quieter footsteps, 2 of 4 slots free.
PrimaryWeapon1: Manticore Executioner, grade 5.
  PrimaryWeapon1: Noise suppressor, Magazine size, Stability, 1 of 4 slots free.
SecondaryWeapon: Manticore Tormentor, grade 5.
  SecondaryWeapon: Stowed reloading, Magazine size, Stability, 1 of 4 slots free.
Modifications are permanent — they cannot be removed or replaced, and a wrong one is recoverable
only by buying and re-upgrading a fresh item.

The suit’s name never comes from Elite’s own localisation. Frontier’s localisation is broken for every suit above grade 1, and it is not an edge case: of 768 SuitLoadout events in a 912-journal corpus, 269 carry an unresolved $UtilitySuit_Class1_Name; token, and every one says Class1 whatever the real class is. Speaking that string would give you the wrong grade more than a third of the time. The symbol plus a shipped table is the only accurate route.

A modification d47 has no name for is said as the symbol Elite wrote, and labelled as one, rather than dropped from the list as though the slot were empty.

What it costs

get_on_foot_engineering

{"type":"object","properties":{"equipment":{"type":"string","description":"A suit or weapon by name \u2014 for example \u0022Maverick\u0022, \u0022Dominator\u0022 or \u0022Karma AR-50\u0022. Defaults to what the Commander is wearing."},"grade":{"type":"integer","description":"The grade to reach, 2 to 5."},"modification":{"type":"string","description":"A modification by name \u2014 for example \u0022Night Vision\u0022, \u0022Magazine Size\u0022 or \u0022Quieter Footsteps\u0022."}},"required":[],"additionalProperties":false}

Name a modification for its recipe and who offers it:

Quieter footsteps — a suit modification.
  Costs: 3 × Settlement Assault Plans (have 0, short 3), 5 × Tactical Plans (have 12),
         5 × Patrol Routes (have 2, short 3), 3 × Micro Hydraulics (have 9),
         8 × Viscoelastic Polymer (have 21)
  From: Yarden Bond.
It is ungraded and permanent, and it needs the item at grade 2 or better before it can be applied
at all.

Or name a suit or weapon and a grade for what reaching it takes:

Maverick Suit to grade 4, at Pioneer Supplies.
  Costs: 4 × Suit Schematic (have 1, short 3), 4 × Health Monitor (have 6),
         4 × Manufacturing Instructions (have 4), 9 × Carbon Fibre Plating (have 30),
         9 × Graphene (have 12)
  Credits from grade 1: 6,150,000.
  Grade 4 carries 3 modification slots. An engineer's base has no Pioneer Supplies, so this has to
  happen before the trip.

The quantities are not the published ones

This is the part most worth knowing about. Every on-foot quantity in the community sources predates the patch that cut them. tools/gen-blueprints.py restates them to what the game actually charges:

Measured against the game itself: 78 ingredient comparisons against the Resources list on 16 real upgrade events, and a zero-remainder cover of 41 material lines across four ShipLocker deltas, each corroborating the other’s leftovers. Every quantity in both tables was observed; nothing is extrapolated. See the finding.

Shipping the published figures would have quoted you two to three times the real cost of everything on foot — and it would have looked right, because it is what every other tool says.

Credits

Reaching a grade costs the item’s grade 1 price times a fixed multiplier — ×4, ×15, ×30, ×50 for grades 2 to 5, per step rather than cumulative, so grade 1 to 5 in full is 99 times the base price. Exact on all 12 weapon upgrades in the corpus across five weapons and four grades, and on both suit grades it covers.

Six of the eleven weapons have no published price, and for those d47 gives no credit figure rather than a total missing a step — the two read identically otherwise.

Where the materials come from

find_micro_resource

{"type":"object","properties":{"material":{"type":"string","description":"A micro-resource by name \u2014 for example \u0022Graphene\u0022, \u0022Circuit Board\u0022 or \u0022Opinion Polls\u0022."},"wanted":{"type":"integer","description":"How many are wanted, for the trade arithmetic. Defaults to the shortfall."}},"required":["material"],"additionalProperties":false}

Graphene — Component. You have 12.
  Found in: Planetary Settlement.
  The Bartender charges 23 barter value for each.
  For 9, you could hand over any of:
    16 × Microelectrode (you have 22)
    23 × Weapon Component (you have 40)
    35 × Chemical Superbase (you have 51)
  Whatever value the last unit does not use is lost, so trade in one go.

The sourcing is more specific than anything a ship material gets. Every micro-resource carries the settlement types, the building types (16 codes, AGRI to STO) and the container types (22 kinds of locker and data port) it turns up in — so “which settlement, which building, which locker” is answerable rather than approximated, where the best a ship material manages is “Planetary Settlement”.

The Bartender’s rate is arithmetic. It takes Components for a fixed barter value each and sells Components for a fixed barter cost, so a trade is floor(Σ(offered × value) ÷ cost) — reproduced exactly on 49 of 49 real trades. This is a better position than the ship material trader, whose rate took a 1,096-trade corpus to pin down.

Two things that change what you should do:

• Leftover value is lost, per trade, with no carry-over. Modelled and refused — crediting the remainder to the next trade drops the score from 49 to 30. So trade in one lump, not in ones.
• Components only. Items and Data are sold for credits and cannot be exchanged at all, and illegal goods are refused outside Anarchy-controlled systems.

The locker cap is per category

1,000 per category, not per item type — the sources conflicted and the game settles it. Across 28,148 locker snapshots, Components total exactly 1,000 in 7,931 of them and never once more, while the largest single component ever held is 94. d47 says so when you are within a quarter of the cap, because gathering past it is wasted time.

Consumables carry a separate cap of 100 per item.

The plan, and your checklist

A suit or weapon plan is the on-foot half of the Loadout tab, and it is the Ships page instantiated against the same drill: an index of what you are wearing and carrying, one item, one slot. It stays a second mode rather than a second tab, because the game separates ship and on-foot hard and so does its vocabulary — but nothing about the layout is redrawn.

What differs is the shape of the thing being planned. A hull’s slot is a place a module goes, and the journal names it. An item on foot has a grade and up to four modification slots, and Elite names none of them — so the slots here are Grade, then Mod 1 to Mod 4, and the numbers are d47’s own. Elite reports what is fitted as a set with no positions in it, so those numbers order your plan and claim nothing about the item.

The grade is the first slot, and that is a routing fact rather than a preference. A grade 1 item has no modification slots at all, and an engineer’s base has no Pioneer Supplies — only an Apex desk. So the plan is ordered the way the trips have to happen.

The plan owns what and the checklist owns when. A plan lives in data/on-foot.json and nothing crosses into your checklist unasked; promoting it produces a proposal you accept, with the grade first and the modifications after it.

Something you do not own is not absent, it is intended. It has no SuitID, so buying it is the plan’s first step rather than a precondition sitting outside it — and buying one adopts the plan rather than making you re-point it.

get_on_foot_plans

{"type":"object","properties":{},"required":[],"additionalProperties":false}

promote_on_foot_plan

{"type":"object","properties":{"item":{"type":"string","description":"The suit or weapon by name. Omit for the suit being worn, or the one weapon carried."},"weapon":{"type":"boolean","description":"True when this is about a hand weapon rather than the suit."}},"required":[],"additionalProperties":false}

drop_on_foot_plan

{"type":"object","properties":{"item":{"type":"string","description":"The suit or weapon by name. Omit for the suit being worn, or the one weapon carried."},"weapon":{"type":"boolean","description":"True when this is about a hand weapon rather than the suit."}},"required":[],"additionalProperties":false}

Dropping a plan keeps whatever it already put on your checklist. You ordered your list around those lines, and removing them silently makes the history wrong.

The file

{
  "kit": [
    {
      "id": "kit-1",
      "equipment": "Maverick Suit",
      "kind": "suit",
      "itemId": 1837009111675068,
      "slots": [
        { "slot": "Grade", "grade": 5 },
        { "slot": "Mod 1", "modification": "Night Vision" }
      ]
    }
  ]
}

id is the plan’s own identity and is independent of itemId from the moment it is created — that independence is what there is to rebind when you buy one. A plan with no itemId is an intended one. A slot that is not Grade or Mod 1 to Mod 4 is refused and reported, because a hardpoint on a suit is a line that could be stored and shown and never promoted.

Hand-edited, it takes effect without a restart, and a line the file gets wrong is reported rather than silently dropped.

Notes for anyone reading the code

Two generated tables sit under this.

tools/gen-onfoot.py builds OnFoot.tsv — suits, hand weapons, hand tools and the modification name map — from EDDiscovery/EliteDangerousCore (Apache-2.0) Items/Suits.cs and Items/HandItems.cs. That is the only permissive source found keyed on the symbols the journal writes: FDevIDs has no suit list and no hand-weapon list, which inverts the ship arrangement where the id list named things and the journal merely agreed.

Names come from there and numbers do not. Its per-grade stats have provenance that varies per figure, one row annotated as a guess, and a transcription bug that assigns the health multipliers to the shield fields. Prices are read off real BuySuit/BuyWeapon events instead, and are blank where none was ever seen.

tools/gen-blueprints.py builds the recipes, with the restatement described above. It refuses to run if an on-foot ingredient size falls outside the measured tables, because one uncorrected number sitting among corrected ones is invisible and is wrong by a factor of two or three.

The modification name map is authored, and that is a deliberate exception. Elite writes weapon_clipsize; every recipe source writes “Magazine Size”; nothing published carries both, and a relaxed match joins only 5 of the 13 symbols the corpus contains. The ship side answered the same problem by refusing to guess, because 786 blueprints is not a table anybody can author. 25 modifications is. Each row records whether its symbol was read out of a real journal or follows the convention and has never been seen, so a join d47 has watched work and one it merely believes in are told apart.