Givmark identity · Edition 02 · October 2026
A mark of
considered trust.
The precision of an assay office.
The method in the open.
The mark
Givmark
A hexagonal perimeter holds a struck diamond: the evidence examined, then marked. Keep the geometry exact and the counter-form clear.
Clear space
Leave at least half the seal’s width around a lockup.
Minimum size
Use the micro seal at 16 to 24px; full geometry from 32px. Wordmark minimum width: 120px.
Keep it precise
No stretching, rotation, extra outlines, gradients or checkmarks within the seal.
16px24px32px48px64px96px
Colour
The Night edition.
One lifted accent.
Teal is the accent and the primary action. It never implies that anything has been checked or approved. Amber always carries a Draft or In development label. Click a swatch to copy its value.
Typography
Manrope / narrative and interface
No mark without evidence.
Headlines 48 to 144px · body 16 to 21px · labels 14px. Restrained weight. Generous space.Fraunces / wordmark and assurance
Givmark. A clearer view.
Keep the serif for the identity and for moments of conclusion.JetBrains Mono / working-paper metadata
GM-WP-DEMO-01
Working-paper references and technical records. Reference numbers shown here are examples only.Status labels
Say what state
it is in.
In developmentDraftExample only
Givmark has issued no marks, reports or opinions. Do not design or show a badge that could be read as a certificate until a published method and a professional basis for issuing it exist.
Voice and evidence
Say what the evidence supports.
Use
“From 2028, Illinois requires the largest frontier AI developers to obtain an annual independent audit.” Cite the source inline.
“Givmark is building independent assurance for AI.”
“Technical testing is done by or with specialist evaluators.”
Avoid
“We audit AI companies.”
“Givmark-certified AI.”
“Nobody else checks AI.”
British English. Declarative. Evidence-led. No emoji, superlatives, implied endorsement or assurance not yet earned.
Components
The system in use.
In development
Draft
Internal controls
Evidence examined
Release approvals, sample of 12.
2028
From 1 January 2028, Illinois requires large frontier AI developers to obtain an annual independent audit.
Latham & Watkins, July 2026Claims and evidence
Safety frameworks, model cards and risk reports make claims. An independent check traces each one to the evidence behind it, or records that the evidence is missing.
Published claims · internal records · test resultsGM-WP-DEMO-01
Example only
Scope: release controls
