Home Technology Solutions Demo About Contact
Interactive Demonstration — Photichain™

Verify it. Then watch it act.

Don't take the claims on faith — run them. In five steps, prove the core of Photichain to yourself, right in this browser:

And an image-block isn’t limited to photographs. The registered image can be data, code, or a file; a transaction or exchange, where a verified match fires an action you design; a medical scan; satellite imagery; or even your existing barcode or QR code, reprogrammed — the code keeps doing its everyday job for every ordinary scanner, while Photichain adds a verified identity underneath

Please note: this in-browser page is a simulation of the core principles — not the enterprise prototype we built and tested. It runs in an ordinary browser tab to make the ideas tangible; the production system is engineered to run as firmware on your existing hardware, closer to the silicon, for far greater speed and scale.
From demo to deployment. This demo shows the principles. The production tool is the Photichain Studio — customized for you after a briefing and set up with our team on a Studio install. That's where you configure what this demo only hints at:
  • Draw the geofence area a record verifies in
  • Bring your own AI computer vision, or use the recognizer we provide
  • Select the sensors that verify your images, and route what happens on a read
  • Choose materials, spectra, and placement (where images are on physical media) — characterized on your own substrates in a live validation test
Request a briefing →
Running edge-local on this device — works even offline

How the demo flows

The core loop — four steps, one protected chain. Registration and reprogramming are both written to the chain of custody; verification and recognition both resolve locally against it. Below the loop: reader gating, field recognition, and challenge tests.

STEP 1 Register image → protected block STEP 2 Program attach / reprogram action STEP 3 Verify exact match / tamper test STEP 4 Recognize camera → action fires ONE PROTECTED CHAIN OF CUSTODY — all steps resolve locally, no server
Run the demo in 60 seconds: 1) Register an image — a bold logo or graphic works best. 2) Tap a quick scenario, then click Program action. 3) Verify the same file in Step 3 and watch the action fire. 4) Run the tamper test — change the image and watch verification flag it. 5) Recognition: show the registered image on another screen (or print it), capture it with your camera, and the action fires on an approximate match. Then reprogram it over the air and run it again.
Step 1 — Register at the point of capture

Create a protected image-block

Choose or capture an image. It is fingerprinted on your device and sealed into the chain below as a new image-block, linked to the block before it.

On a phone, this lets you take a photo or pick from your library.
Step 2 — Program & reprogram

Attach mission logic to the image-block

Every registered image carries a programmable action that fires on a verified match. Change it any time — the image-block is re-tasked without re-registering, and the change is recorded in the chain of custody.

Step 3 — Verify (integrity mode)

Authenticate against the chain

Present an image. It is checked against the protected chain locally, with no server call. A registered, unaltered image passes and its programmed action executes; anything altered or unknown fails. The few milliseconds shown is browser overhead on a general-purpose CPU — a ceiling, not the target. In firmware on a DPU it is engineered to run far faster.

Step 4 — Tamper test

Alter one pixel. Watch it fail.

This takes an image you registered, changes a single pixel, and runs the same verification. The altered copy no longer matches its protected image-block — tamper-evidence in action.

Step 5 — Reader gating

Reading is a privilege: only assigned devices can decode

A protected image-block is cryptographically unreadable to unassigned devices — you decide which hardware holds the key. Toggle provisioning below (simulated for this demo), then attempt a read as this browser: first without the key, then with it. The anti‑enumeration consequence matters as much as the security one — to everything you have not assigned, your protected assets yield nothing: no identity, no data, no inventory intelligence.

This browser acting as Dock Camera A
Fleet camera F-3 remote reader · simulated
Step 6 — Recognition mode (camera)

See it in the field: point your camera at a registered image

Display one of your registered images on another screen — or print it — then point your camera at it. Photichain recognizes the genuine, registered image and resolves it. A photo of a photo isn’t the sealed original, so it doesn’t carry the same proof. What you assign is what verifies.

On a phone, tap to take a photo or pick from your library.
Try to fool it

A demo that only succeeds is a demo you shouldn’t trust

Verification here is two-layered by design: an exact layer (Step 3) where changing a single bit fails, and an approximate recognition layer (Step 6) that tolerates real capture noise while rejecting different content. These challenges show where each layer draws its line — run them and watch the system explain its own boundaries.

  • Crop challenge: crop one of your registered images and verify the crop in Step 3 — the exact layer fails (a crop is a different file), while recognition in Step 6 may still identify what it sees. That split is the two-layer design working, not a contradiction.
  • Screenshot challenge: screenshot a registered image and verify the screenshot — exact fails (different bytes), recognition recognizes. Provenance and recognition, separated on purpose.
  • Different-content challenge: point the camera at anything you never registered — recognition stays below threshold and no action fires.

The chain of custody

Each block below is sealed against the one before it. Registrations and reprogramming events are both recorded, so re-tasking an image-block never breaks the custody trail.

From this demo to your hardware

What you just used is the floor, not the ceiling. Every capability on this page scales up in the production architecture:

Today: browser tab

Tomorrow: your silicon

This page runs in a sandboxed browser. The production architecture is resident on the hardware you already own — CPUs, DPUs, and SmartNICs in the field, or servers and workstations inside your own network — and can equally be designed into hardware you are still developing, so it ships native rather than retrofitted. Either way, it verifies at the point of capture.

Today: simplified matching

Designed for: AI-driven recognition

Recognition here is deliberately simple. The patented architecture specifies AI-based computer vision — built to recognize registered images through the lighting, angles, and distances of real field conditions.

Today: one device

At scale: your whole fleet

You reprogrammed one image-block over the air. The same patented capability is designed to re-task verification logic, identities, and triggered actions across an entire deployed fleet — with no recovery, no replacement.

Beyond this page

The enterprise prototype

This browser demo is the floor — the simplest public expression of Photichain. Response Images has also built and exercised an enterprise-scale prototype on enterprise cloud infrastructure, running one-way (point-of-capture) and two-way (round-trip) verification pipelines — the program behind the consensus-free architecture presented across this site. Full instrumented benchmarking is the current milestone.

To arrange a technical briefing on the prototype program — or to explore developing the capability for your systems and sector — get in touch.

Request a technical briefing
The same engine, digital

This demo shows an image. It works the same on a file.

What you just did to an image-block, Photichain does to any digital object — one engine, edge and digital. Instead of a captured image, seal a transaction, a document, a record, or a data feed the moment it’s created, and verify it locally as it moves between systems, data centers, or regions.

Digital use case

Transactions & exchanges

Confirm a transaction, claim, or handoff is genuine and unaltered as it happens — both sides checked locally, with a record kept when you want one.

Digital use case

Documents & records

Seal a contract, scan, manifest, or evidence file and know at every step it’s the same one — change a character and it shows.

Digital use case

Data on the move

When data travels between sites or across the world, confirm it arrived exactly as it left — checked at both ends, no trip to a middleman.

Most sectors use both: capture at the edge and verify digital objects in motion. See the digital page for the full picture.

About this demonstration. This page is the simplest possible expression of Photichain — a concept demonstration of the patented principles — point-of-capture registration, tamper-evident image-blocks, programmable and reprogrammable actions, closed-loop verification, and edge-local resolution — running entirely in your browser. It is intentionally not the firmware-resident production implementation — that architecture is designed to run on dedicated hardware, closer to the silicon, under AI-driven recognition. Timing shown is measured live on your device for this demo and is not a benchmark of the production system. The demo ledger lives in this page's memory only: nothing is uploaded, and refreshing the page clears it. Integrity mode matches the exact registered file, so a re-saved or re-compressed copy counts as altered. Recognition mode uses approximate visual matching computed on your device — results vary with lighting, angle, and framing — and no AI model or server is involved in either mode.

© Response Images LLC. Photichain is patented technology — U.S. Patent 12,395,313. This interactive page demonstrates core principles in-browser and makes no performance claims beyond timings measured live on your device.

Photichain™ is a trademark of Response Images, LLC. What is shown here reflects one embodiment and is not limited to the embodiments above; it does not represent the boundary of U.S. Patent 12,395,313 B1.