A live, experimental demo showing how our core technology, BlkBolt, can generate and verify signature artifacts for digital content and assets using machine-learning models instead of traditional keys or certificates.

This demo is designed to be safe, non-destructive, and educational — not a production security tool. By participating in this demo you are actively helping us conduct research in a real-world situation. Thank you in advance!

Note: We use the term “signature” intentionally, because the goals—authorship, provenance, integrity, and verification—are familiar, even though the mechanism is different.

It demonstrates one way to create a verifiable signature artifact for a digital file.

Specifically, it shows how you can:

• Prove a file existed in a specific form
• Record who created the signature and when
• Verify the file hasn’t been modified or the signature artifact wasn't tampered with
• Revoke or invalidate that proof later if needed

All without storing your original file or relying on long-lived private keys.

Not in the traditional cryptographic sense.

This demo uses model-based verification, where authenticity is tied to a specific trained model and lease, rather than a static private key.

Think of it as a signature artifact for authenticity and verification, not a legal signature.

What actually happens when I upload a file?

• Your file is hashed locally on the server
• Only the hash and minimal metadata are processed
• A leased model encodes that information into a verification artifact
• Only the model used to create the signature artifact can decode it
• The original file and signature artifact are returned to you as a package
• You control the package and signature artifact — nothing is retained on our side

Your file is not stored after the request completes.

The package contains:

• Your original file
• A small README text file outlining context and scope
• A signature.json file containing...
• A request ID
• A model ID for the model used to generate the artifact
• A base64 encoded signature artifact

No private keys, secrets, or stored files are included.

Yes, and we encourage it! You can upload a package to be verified here in order to...

• Confirm the original file hasn't changed
• Inspect decoded metadata (e.g. create timestamp, etc.)
• Validate the lease is still active

No.

Files are processed in memory for hashing and packaging, then immediately discarded. This demo, and system in general, is intentionally designed to avoid persistent storage of user data.

You’re welcome to try 🙂. In fact, we'd love for you to try, especially if you have access to a quantum computer. This demo is intentionally designed to be inspectable, shareable, and challengeable.

You’re encouraged to:

• Modify files after signing
• Alter metadata in the package and signature.json file
• Reuse signature artifacts after a lease is expired or revoked
• Attempt to forge, replay, or reverse engineer signatures
• Explore edge cases and failure modes

If something breaks, that’s valuable signal — not a problem. There’s no penalty for experimentation, and no sensitive data involved. Every successful or failed attempt helps inform future design decisions.

If you think you've got it, email us at support@lyfe.ninja with the worked "cracked" in the subject.

A lease is a temporary right to use a specific model for encoding.

In this demo:

• Leases are short-lived lasting only 15 minutes
• A model can only have one active lease
• Only one user can use a lease at a time
• Leases can be revoked or expire automatically
• New leases may use the same or different model

Leases allow:

• Revocation by design
• Isolation between users
• Limited exposure if an signature artifact is shared
• Safer experimentation

Once a lease is revoked or expires, verification behavior changes.

After revocation or expiration:

• New signature artifacts cannot be created
• Existing signature artifacts may fail verification
• The system visibly demonstrates revocation effects

This is intentional and part of the demo.

Technically, deleting the underlying model permanently invalidates all signature artifacts created by it. However, given the time and resources required to train these models, that would be wasteful behavior, so leases are a great practical alternative.

This demo is not a cryptographic claim.

It explores model-based isolation and verification concepts that may be useful in future systems, including post-quantum contexts — but no formal guarantees are made here.

Want to help us research this topic? Fill out our Work With Us form or email us at connect@lyfe.ninja.

This demo is especially interesting for:

• Developers and security engineers
• Researchers exploring alternative data encoding, attestation, and verification methods
• Creators curious about authenticity and provenance (even AI creators)
• Anyone interested in experimental security models

No — and that’s the point. This is a live research demo meant to:

• Explore ideas in the open
• Gather real-world feedback
• Test assumptions before productization

Sorry, if it's a little slow, but speed will come later.

After completing certain stages of the demo (either generating or verifying a signature artifact), you’ll be invited to share feedback. You can also provide feedback immediately here . Please contact us if you’re interested in:

• Strategic or design partnerships
• Research collaboration
• Early access to future experiments

Reach out directly by filling out our Work With Us form or emailing us at connect@lyfe.ninja.
Follow us on LinkedIn.

No. Please upload only non-sensitive files you’re comfortable experimenting with. This demo is for learning and exploration only.