Untokened receipt markUNtokened
UNTOKENED / BRAND REFERENCE V 1.0

Know what every
prompt costs.

The design system for a gateway that makes AI usage legible. Every call logged. Every model priced. No guesswork in the numbers.

DOCUMENT 001IDENTITY / SYSTEM / VOICE6 SECTIONS
/01

Identity

A receipt, reduced to its essentials. The crossed token is a small statement of intent: less spent, more accounted for.

PRIMARY WORDMARK
Untokened receipt markUNtokened
UN / BOLD   tokened / REGULAR
MARKUntokened receipt markRECEIPT / TOKEN
FAVICONUntokened favicon48 × 48
/02

Color

Paper and ink do the heavy lifting. Green marks savings. Amber asks for attention. Red means the budget has been crossed.

light themePAPER / INK
Paper#F6F4EE
Surface#FFFFFF
Ink#141414
Muted text#6B6A65
Rule#DDDAD0
Saved#12A150
Signal#E8A317
Redline#D7372F
dark themeNIGHT / INK
Paper#121211
Surface#1B1B19
Ink#EDEBE4
Muted text#9A988F
Rule#2C2B28
Saved#12A150
Signal#E8A317
Redline#D7372F

Chart palette

SIX SERIES / COLOR-BLIND CONSIDERED
01#0072B2
02#D55E00
03#009E73
04#CC79A7
05#56B4E9
06#E69F00

Text-safe accent variants are used on light surfaces where the exact signal colors would fall below AA contrast. The original colors remain unchanged in the palette.

/03

Typography

Geist for the human layer. Geist Mono for anything the machine can count, identify, or charge for.

01 / INTERFACE

Geist

Headlines, labels, and explanatory text.

02 / DATA

Geist Mono

Costs, tokens, models, IDs, and code.

56 PXThe total adds up.
40 PXKnow your costs.
28 PXEvery request, accounted for.
20 PXUsage by model
16 PXAll requests are logged and priced.
14 PXLast 30 days · 12,418 requests
12 PXRECEIPT ID · req_8f29d1

Numbers are tabular

$0.0042  /  12,418

0123456789

Spacing & shape

4 PX BASE UNIT
4
8
16
24
32
48
6 PX / CONTROLS
10 PX / CARDS
1 PX / RULES
/04

Components

The working parts. Quiet until there is something worth pointing out.

Buttons

DEFAULT / SECONDARY / GHOST / DESTRUCTIVE

Inputs

NORMAL / FOCUS / ERROR / DISABLED

Select

PROVIDER FILTER

Tabs

ACTIVE / INACTIVE / DISABLED
12,418 requests in this period.

Badges & status

LABELS / SYSTEM STATES
Saved
Draft
Active
Over budget
HealthyNear limitOver budget

Model chips

PROVIDER + MODEL ID
gpt-4.1-minigemini-2.5-proclaude-sonnet-4

Stat tiles

LABEL / NUMBER / PERIOD DELTA

Total spend

$284.62

↓ 18.4% vs. last period

Requests

12,418

↑ 7.2% vs. last period

Table & cost cells

MONO COSTS / RIGHT-ALIGNED
ModelStatusTokensCost
gpt-4.1-miniLogged1,842$0.0042
gemini-2.5-proNear limit3,206$0.0178
claude-sonnet-4Over budget2,914$0.0314

Empty state

ZERO DATA

No requests yet.

Requests will appear here after the first API call.

Code block

COPYABLE REQUEST
REQUEST · bash
curl https://api.untokened.com/v1/chat \
  -H "Authorization: Bearer $API_KEY"

Dialog

CONFIRMATION

Toast

FEEDBACK
/05

Voice

Plain. Exact. A little dry. Use numbers instead of adjectives. Say what happened, then what to do.

01 / HEADLINE

“Know what every prompt costs.”

Lead with the outcome, not the technology.

02 / INSIGHT

“You spent $284.62 this month. That's 18.4% less than last month.”

Put the number first. Explain the comparison.

03 / WARNING

“You've used 92% of your budget. $24.00 remains.”

State the limit without creating drama.

NOT THIS: “Revolutionary insights that supercharge your AI journey.”