LIVE PRODUCTION SERVERS

Break This

100,000 TREE BOUNTY

Can you break the Three-Layer Fortress? Every test below hits live production servers. No API key required. No sandbox. This is real.

01

Cross-Server Round Trip

Challenge
Encrypt on Server A, decrypt on a different Server B. If the data survives the cross-mesh hop intact, the shared provenance model is working.
02

The Wall — Haiku Steganography

Challenge
Same SSN ("123-45-6789") encrypted 5 times. Each output is wrapped in different haiku poetry. Can you tell which haiku contains the SSN?

The steganographic layer wraps every encrypted payload in natural-language poetry. Five encryptions of the exact same data produce five completely different poems. You cannot tell which haiku hides what.

03

Live Server Mesh Status

TreeChain operates on a 5-server mesh across 3 continents. Each server independently encrypts using shared provenance via MongoDB Atlas. Hit all 5 with the same plaintext and watch the outputs diverge.

🇫🇮
EU-Helsinki
api-eu.treechain.ai
--
🇺🇸
US-Oregon
api-us.treechain.ai
--
🇸🇬
APAC-Singapore
api-apac.treechain.ai
--
🇺🇸
US-Ashburn
178.156.228.56:8001
--
☁️
Render
glyphjammer-api.onrender.com
--
04

Determinism Test

Challenge
Encrypt the same plaintext N times. If even one output repeats, the system is deterministic and vulnerable to pattern analysis. P(collision) should = 0.
Collision Probability = n(n-1) / (2 x |Omega|)
where |Omega| = output space size, n = number of encryptions
05

Known Plaintext Attack

Attack Model
Given 10 plaintext-ciphertext pairs {(P1,C1), (P2,C2), ..., (P10,C10)}, can the adversary predict C11 for a known P11?

Classic cryptanalysis attack. We give you 10 known pairs, then challenge you to predict the next. The Stochastic Rotor makes this impossible.

06

Corpus Analysis — Match the Pairs

Challenge
Enter multiple messages. All get encrypted and shuffled. Can you match each encrypted output back to its plaintext? Without the key, this is a random assignment problem.
07

Haiku Boundary Problem

Challenge
5 poems below. Some contain encrypted data, some are real poetry. Guess which is which. Steganographic indistinguishability means you cannot tell.

We mix real haiku with TreeChain-generated steganographic wrappers. If you can reliably distinguish them, the steganographic layer is broken. Click each card to mark it as "encrypted" (red) or "real poetry" (green).

08

Frequency Analysis (Chi-Squared)

Null Hypothesis (H0)
Glyph frequencies follow uniform distribution: f(g) = 1/|G| for all glyphs g in set G.
Alternative Hypothesis (H1)
Glyph frequencies deviate from uniform: exists g : f(g) != 1/|G| (detectable patterns exist)
X2 = Sum( (Oi - Ei)2 / Ei )
Oi = observed frequency, Ei = expected frequency (n/k), df = k-1
09

Rotor Animation — Position-Dependent Mapping

Visualization
Like an Enigma rotor: the same input character at different positions maps to different glyphs. Watch the position-dependent substitution change as the rotor advances.

Rotor Position: 0 | Input: A | Same char, different output at every position

Bug Bounty Program

Tier 1: Convergence

Find a collision — prove the same plaintext always produces the same ciphertext

10,000 TREE

Tier 2: Leakage

Extract PII from haiku wrapper without encryption key

50,000 TREE

Tier 3: Ghost Break

Reverse the Glyph Rotor mapping to reveal ChaCha20 bitstream

100,000 TREE
Submit findings: security@treechain.ai

Download Sample Data

Generate encrypted datasets for offline cryptanalysis. Each sample includes plaintext, glyph output, and server metadata.

1K Samples

1,000 encrypted records in JSON format (~2 MB)

10K Samples

10,000 encrypted records in JSON format (~20 MB)