SentriAI

🛡️ SentriAI — The Decision Integrity Layer for Autonomous Finance

“Every major DeFi failure began with a decision no one could independently verify.”

SentriAI makes that impossible.

SentriAI is a deterministic, autonomous compliance agent built natively for the Amadeus blockchain.It transforms AI-driven financial decisions from opaque, off-chain judgments into verifiable, permanent on-chain facts.

As autonomous agents begin to trade, lend, and allocate real capital, trust can no longer depend on private servers, unverifiable heuristics, or “trust me” AI outputs.SentriAI ensures that every decision an agent makes can be replayed, audited, and proven—forever.

🛑 The Problem

DeFi infrastructure is institution-ready.

AI decision-making is not.

Today:

• Compliance logic runs off-chain inside private systems

• Decisions cannot be independently reproduced or verified

• There is no permanent record explaining why an action was allowed or blocked

• Autonomous agents execute high-impact actions with no auditable reasoning trail

This creates a fatal gap:

You can verify transactions — but not the decisions that caused them.

As agent-driven finance scales, unverifiable AI decisions become systemic risk.

✅ The SentriAI Solution

SentriAI is a deterministic decision-verification agent that evaluates wallets and transactions using transparent, reproducible logic — then anchors the result as a permanent on-chain artifact.

Every compliance decision becomes:

• Deterministic — same input → same output

• Auditable — logic and results can be independently recomputed

• Immutable — decisions are stored as permanent records

• Machine-consumable — usable directly by other agents and protocols

Compliance stops being a black box.It becomes shared infrastructure.

🧠 How It Works (High Level)

1. A wallet or transaction is submitted for evaluation

2. SentriAI executes three deterministic agents:

   • Sanctions Risk Agent

   • Behavioral Risk Agent

   • Reputation Risk Agent

3. Scores are combined using a transparent, reproducible formula

4. A final verdict is produced: APPROVED / WARNING / BLOCKED

5. The full decision payload is deterministically hashed

6. The hash is anchored as a permanent provenance record

The result can be verified forever — without trusting SentriAI itself.

🧾 Deterministic AI Provenance (Arweave Bonus)

SentriAI treats AI decisions as first-class artifacts, not ephemeral outputs.

For every compliance check, SentriAI generates:

• A deterministic JSON payload of the AI computation

• A cryptographic hash of that payload

• A deterministic Arweave transaction ID derived from the hash

This creates an Eternal Decision Ledger, where:

• The same input always produces the same record

• Anyone can recompute and verify the decision

• AI decisions become permanent, auditable facts

This end-to-end provenance pipeline directly aligns with the Best Provenance Architecture challenge.

🧩 Architecture Overview

Data Layer

• Oracle streams (sanctions lists, behavioral signals)

• Historical risk metadata

Agent Layer (WASM Runtime)

• Deterministic compliance agents

• Reproducible scoring logic

• No randomness, no time dependency

Settlement & Provenance

• State proofs for decision integrity

• Deterministic hashes

• Permanent provenance records anchored to Arweave

⚙️ What Works Today vs. Future

Works Today

• Deterministic agent logic

• WASM-compatible execution model

• State-proof-based audit trails

• Provenance anchoring architecture

• Honest, reproducible decision-making

Future (Amadeus Roadmap)

• uPoW-backed autonomous enforcement

• Consensus-level agent execution

• Privacy-preserving compliance proofs

SentriAI is designed to work now, and scale naturally into the future.

💰 Monetization Path

SentriAI follows a clear B2B infrastructure model:

• Protocols pay for verified compliance decisions

• Pricing scales with transaction volume

• Designed for institutions, DAOs, and agent-driven protocols

This mirrors how compliance infrastructure is monetized in real financial systems today.

🌍 Why This Matters

As autonomous agents begin to:

• Trade

• Lend

• Allocate capital

• Coordinate with each other

They require a shared source of truth for trust.

SentriAI provides that foundation:

A world where AI decisions are verifiable facts, not opaque guesses.

If agents are the actors of the autonomous economy, SentriAI is the decision integrity layer that makes them safe to trust.

🏁 Why Amadeus

SentriAI is built for Amadeus because Amadeus is built for agents.

• Deterministic execution via WASM

• Verifiable state transitions

• Native support for future uPoW compute

• On-chain intelligence as a protocol primitive

SentriAI is not just an application on Amadeus —it is core infrastructure for the autonomous economy.

🔒 SentriAI

Every decision verified.

Every action remembered.

Forever.

🔗 Live Demo: https://sentri-ai-hazel.vercel.app/

🔗 GitHub: https://github.com/pragyanthapa/SentriAI

🎥 Video: https://youtu.be/zPymimAmuxs