···················
Proof-of-life protocol · v1.0

Prove reality,
don't detect fakes.

PRAMA is a cryptographic alibi engine that mathematically proves where you were, when you were there, producing court-admissible evidence that survives hostile legal scrutiny.

0
Bytes of plaintext stored
2%
Max daily battery drain
~3ms
Per snapshot chain hash
The Problem

Detection is losing the war.
Provenance has to start at capture.

The forensic stack we inherited assumes you can spot fakes after the fact. That contract is broken. Three independent failures push the legal and social burden back onto the subject of the lie.

01 · C2PA · Authenticated Fakes

Metadata washing

C2PA's signature scope explicitly excludes GPS and timestamps so files can be re-encoded. An attacker can rewrite where and when a verified image was captured while the cryptographic signature stays valid.

02 · GPS Spoofing · $5 emulator

Location is a suggestion

Software GPS emulators cost the price of a coffee. Carrier ping logs are subpoena-friendly. Today's defense, "my phone says I was here," collapses the moment opposing counsel reproduces the same coordinates on a Raspberry Pi.

03 · Deepfakes · Detection drift

Reactive forensics is losing

Generative models re-train on every detector. Pixel-level forensics is an arms race with diminishing returns. The only durable answer is to invert the proof: stop arguing about the fake, prove the reality.

How it works

Silicon to courtroom,
in four steps.

PRAMA runs as an invisible background protocol on your phone. Each layer adds a different proof: that the data came from real silicon, from your body, and from a moment in time that can never be rewritten.

  1. Step 01

    Capture.

    A quiet record running in the background

    Your phone takes a snapshot whenever something changes. You unlock the screen, your cell tower swaps, your watch logs a heartbeat. No prompts, no timers, just a passive ledger of where you actually were.

    Capturing signals

    • Location

      38.9954°N · 76.93°W

    • Wi-Fi

      14 BSSIDs nearby

    • Bluetooth

      9 devices

    • Cell tower

      T-Mobile · LTE 66

    • Altitude

      142 m

    • Heart rate

      68 bpm

    Hashed locally, never transmitted

  2. Step 02

    Bind.

    Tied to you, on your real device

    Each snapshot is signed by Face ID and a private key that only your phone holds, then certified by Apple App Attest or Google Play Integrity. The frame is provably yours, from a real device, signed by you.

    Identity bound

    • Face ID

      Verified

      this unlock
    • Heartbeat

      68 bpm

      paired watch
    • Real device

      App Attest

      passed
    • Your signature

      Ed25519

      stamped

    Signed inside hardware, never exported

  3. Step 03

    Seal.

    Encrypted on device, by you

    Your raw data is encrypted on your phone using a key only you hold. The server only ever sees a sealed blob and a unique fingerprint. The key stays with you, always.

    Sealed on device

    Your fingerprint

    0xC4 92F1 7AB8 9D7E

    • Your data

      Encrypted on your phone

    • Your key

      Stays on your phone, always

    • Server sees

      The sealed blob and fingerprint only

    Only you can open your snapshot

  4. Step 04

    Anchor.

    Permanent, public, verifiable

    Every few hours your fingerprint is published to a public ledger with a verified timestamp. From that instant the snapshot is mathematically frozen. Anyone with the receipt can verify it independently, years later.

    Anchored

    Locked in time

    Permanent record

    • Time17:08:31 UTC · 24 May 2026
    • LedgerArbitrum public blockchain
    • Verified3 independent time servers
    • ReceiptTX 0x4F3A…92E1

    Anyone can verify independently

Architecture

Signals become context.
Context becomes proof.

GPS lies. A single radio is forgeable. A ring of independent radios, glued together by a biometrically signed Merkle root, is physically expensive to fake, and trivially cheap to verify.

  • 12 random Wi-Fi BSSIDs per snapshot
  • 4 Bluetooth proximity beacons
  • ENF spectral hash from grid hum
  • Roughtime-verified timestamp ±42 ms

Nearby Networks

Live
  • xfinitywifi

    94:b3:4f:51:11:61 · -52 dBm

  • ATT-WiFi-9F2C

    3c:37:86:2a:b1:04 · -58 dBm

  • NETGEAR47-5G

    a0:40:a0:1d:7e:22 · -64 dBm

  • Spectrum_Mobile_4423

    94:b3:4f:50:f7:90 · -71 dBm

  • TP-LINK_E2C4

    94:b3:4f:50:da:82 · -78 dBm

Hashed locally · never transmitted

Court-admissible by construction

Field cases · The proof gap

These are not
hypothetical.

Six real people, on the record: jail over unverified AI texts, a deepfake by someone trusted, a live likeness stolen from Zoom, imposter accounts built from a meteorologist's face, non-consensual deepfake imagery of a broadcast journalist, and fans targeted simply for supporting a creator they follow. Every card has two faces. The voice. And the receipt that would have changed the outcome.

ABC7 ChicagoJan 2026

I end up getting arrested for violating my bond. No one verified the evidence. It was horrific. Why would someone knowingly send a text from their phone that could get them arrested? People are not really that stupid.

Melissa SimsArrested twice · AI-generated texts as evidence · Florida
View source
BBC · File on 42024

I was screaming and crying and violently scrolling through my phone to work out what I was reading and what I was looking at. I knew that this could genuinely ruin my life.

JodieDeepfake victim · fabricated by someone she trusted · UK
View source
Joshua Davis · Substack2024

They took Zoom footage from a past session and used AI to animate me doing something I never did. They used my likeness, my actual face and body, and inserted it into a real time fake video stream. This isn't just misleading. It's manipulative. It's dangerous.

Joshua DavisPrivate individual · live AI likeness theft
View source
Yahoo · Tennessean2025

Last fall, fake Facebook accounts pretending to be me started popping up. These accounts used fake pictures that showed my face on someone else's semi-nude body. They also make fake videos that used my face and my name to try to convince people that it was really me. Discovering these imposter accounts and seeing the degrading fake images and videos was devastating to me.

Bree SmithFormer TV meteorologist · identity deepfakes · Tennessee
View source
Channel 4 · AOL2024

It feels like a violation. It just feels really sinister that someone out there who's put this together, I can't see them, and they can see this kind of imaginary version of me, this fake version of me. You can't unsee that.

Cathy NewmanBroadcast journalist · non-consensual deepfake imagery · UK
View source
Sportskeeda · Twitter/XMar 2023

I'm so sorry to my community members that have had deep fakes made of them just for supporting me. You shouldn't have had to experience that in any way. please protect each other, especially the minors! Report those accounts and just know it says more about them, not you. Stay safe.

ValkyraeYouTube creator · fans deepfaked by association
View source

Six different attacks. Same missing piece: proof that what they say happened, actually happened.

PRAMA is that proof, captured before the attack.
Cross-examination · Q.A

Everything you
need to know.

Five exhibits. Nineteen questions. The same answers PRAMA gives any judge, lawyer, or opposing party who walks into the room.

Exhibit 01 · Privacy & data
  • A.

    No. Your location data is encrypted on your device before it ever leaves your phone. Our servers store only encrypted data that we mathematically cannot read. Not now. Not ever.

Closed alpha · Q3 2026

You are here.
Provably.

The first wave is invite-only and capped at 500 devices. Tell us who you are and where you operate. We route invites jurisdictionally based on legal-admissibility maturity.

Read the blueprint

Zero-knowledge by default · we store only your email