Press kit

Quotable facts and the brand mark.

Everything an analyst, journalist, or podcast host needs to describe TokenForensics correctly, with one click. No logo grid of fictional partnerships, no boilerplate.

THE MARK

Brand mark.

Three corner-linked rounded squares on a descending diagonal — the hash chain of TokenAuditEvent blocks. Monochrome by design; renders in currentColor so it adapts to light and dark backgrounds.

TokenForensics.ai

On light · currentColor renders the mark

TokenForensics.ai

On dark · currentColor renders the mark

The source is src/components/BrandMark.tsx on GitHub — Apache-2.0 with the rest of the repository.

THE TYPE

Wordmark and colors.

Wordmark: TokenForensics.ai in Geist Sans, font-semibold, tracking-tight. The .ai stays lowercase. Minimum clear space around the lockup is the height of one block of the mark.

Palette: Tailwind zinc plus pure black/white. One accent — Evidence Amber(amber-500, #f59e0b) — reserved for forensic significance (an IOC match, a tampered chain link, a flagged span). Never decorative.

Full brand kit: docs/brand.md.

THE FACTS

Quotable copy.

One-line
TokenForensics is the investigation layer above AI infrastructure — chain-of-custody token-level capture, replay, and attribution for production LLM and agent systems.
Category
AI forensics — sits downstream of any AI gateway, complementary to any AI observability tool. Never a competitor to either.
What's shipped
Pillar 1 (tamper-evident capture): cryptographically signed, hash-chained, daily-Merkle-rooted token-level capture across OpenAI, Anthropic, and the major agent frameworks. Pillars 2–8 are in plan.
Open standard
TFEF — Token Forensics Exchange Format — is the vendor-neutral schema the substrate is built on. Published openly; the long-term moat is schema ownership, not feature parity.

THE LINKS

Where to look next.

CONTACT

Reach the team.

For press inquiries, partnership questions, or to be added to the design-partner program:

scott@promptreports.ai