Now in Early Access · Trusted by Finance Teams

The AI Compliance Layer That Governs Every Rupee Spent.

AegisOS Comply intercepts every AI-driven spend, evaluates it against your policies in milliseconds, and creates an immutable audit trail — with SOX-ready exports, real-time analytics, Ed25519 agent signing, and an AI-powered policy generator built in.

Get Started Free → See How It Works

Intent Decision

₹1,200

Office Supplies · Procurement Agent

Approved

Matched: Default Allow · 4ms

Pending Approval

₹75,000

Vendor Invoice · Finance review

Pending

Matched: Medium-Value Rule · 6ms

Auto-Denied

₹6,00,000

Exceeds limit policy

Denied

Matched: High-Value Deny · 3ms

The Problem

AI agents spend money. Nobody's watching.

As AI models autonomously trigger financial transactions, your organization is exposed — no visibility, no governance, no audit trail.

🔓

No Access Controls

AI agents execute payments without any policy guardrails or spend limits.

👻

Zero Audit Trail

Finance teams can't reconstruct what the AI did, when, or why — audits become impossible.

⚖️

Compliance Gaps

RBI, EU AI Act, SOX, and SOC 2 all require documented AI decision trails. Most teams have none.

Our Solution

Governance built in, not bolted on.

AegisOS Comply sits between your AI agents and payment infrastructure — evaluating every spend in real time.

🛡️

Real-Time Policy Engine

Every spend intent is evaluated against your rules in under 10ms before any money moves.

🔗

Immutable Audit Chain

SHA-256 checksum-chained logs ensure your audit trail cannot be tampered with — ever.

✅

Human-in-the-Loop

High-value or uncertain transactions are routed for human approval before execution.

The Flow

From intent to decision in milliseconds.

Every AI spend request flows through a deterministic, observable pipeline — with full traceability at each step.

🤖

SDK Call

Agent submits intent via Node.js or Python SDK

📨

Received

API returns 202 Accepted instantly — non-blocking

⚙️

Policy Engine

Rules evaluated against amount, confidence, risk

📋

Decision

Allow, deny, or route to human approver

🔒

Audit Logged

SHA-256 chained entry written — immutable forever

✅

Auto Approved

Under ₹50K · High confidence · Low risk

🙋

Human Review

₹50K–₹5L or low confidence score

🚫

Auto Denied

Over ₹5L or blocked merchant category

What We Provide

Everything to govern AI spending.

A complete platform — SDK to dashboard to compliance exports — built for finance and engineering teams.

🧠

Policy Rule Engine

Define rules on amount, currency, merchant category, agent risk level, time of day, and AI confidence. Supports AND/OR logic with priority ordering.

allowdenyrequire_approvalpriority

🔗

Immutable Audit Trail

SHA-256 checksum-chained logs. Each entry references the previous — making retroactive tampering cryptographically detectable.

SHA-256Append-onlyTamper-proof

🙋

Smart Approval Routing

JSON routing rules auto-escalate high-value or risky transactions to the right role or user — with expiry timers, approval history, and Slack / email notifications.

Auto-routeSlackExpiry

📦

SDKs — Node.js & Python

Zero-dependency TypeScript SDK and Python SDK. Drop into any AI agent in minutes. Full type safety, idempotency support, and error handling.

TypeScriptPythonZero deps

📊

Compliance Export Engine

Export audit trails as JSON or CSV in RBI, EU AI Act, SOX, or generic formats. SOX export includes control references, system-generated flags, and approval status.

RBIEU AI ActSOXCSV

🏢

Multi-Tenant Architecture

Full org isolation with custom RBAC — departments, teams, and 5 system roles. Each org has its own agents, policies, API keys, and audit logs.

Custom rolesDepartmentsTeams

🔑

Ed25519 Agent Identity

Cryptographically sign every SDK request with per-agent keypairs. Nonce and timestamp validation prevents replayed or spoofed intent submission.

Ed25519Replay protectionCryptographic signing

💱

Multi-Currency Normalization

Normalizes spend amounts across USD, INR, EUR, and GBP in real time with intelligent caching and automatic fallback rates — policies always evaluate in a consistent base currency.

Live FX ratesAuto-fallbackMulti-currency

🔬

Policy Simulator

Test a new policy against 30 days of real historical intents before activating it. Detects shadowed rules and conflicting policies with existing active rules.

SimulateConflict detectionImpact %

🤖

AI Policy Generator

Describe your compliance need in plain English — Claude generates a validated policy draft ready for review. From "block weekend spend over ₹10K" to a working rule in seconds.

Claude AINatural languageAuto-validate

📈

Spend Analytics

Live KPI dashboard with spend trends, approval funnel, agent-level breakdowns, and forecast modeling — giving finance teams full visibility into where AI money goes.

Live dashboardForecastAgent breakdown

⚠️

Risk Engine

Real-time risk scoring per agent based on spend patterns, denial rate, merchant categories, and anomaly detection. High-risk agents are flagged and wallet-blocked automatically.

Risk scoreWallet blockAnomaly

Developer SDK

Integrate in 5 lines of code.

Zero dependencies. Full TypeScript types. Works in Node.js and Python.

agent.ts

1import { AegisClient } from '@aegis-os/sdk';
2
3// Initialize once — reuse across your agent
4const aegis = new AegisClient({
5  apiKey: process.env.AEGIS_API_KEY,
6});
7
8// Before any AI spend, submit an intent
9const intent = await aegis.intents.submit({
10  agentId:    '653b13da-6ebf-41f2-abea-2688e9aefb5f',
11  amount:     1200,
12  currency:   'INR',
13  reason:     'Purchase office supplies',
14  confidence: 0.95,
15  merchant:   'Amazon Business',
16});
17
18if (intent.status === 'approved') {
19  executePayment(intent.id);          // ✅ proceed
20} else if (intent.status === 'pending_approval') {
21  console.log('Awaiting finance team approval…');
22} else {
23  console.error('Denied:', intent.policyDecision); // 🚫 stop
24}

agent.py

1from aegis_os import AegisClient
2
3# Context manager — auto-closes session
4with AegisClient(api_key="aegis_sk_…") as aegis:
5
6  intent = aegis.intents.submit(
7    agent_id    = "653b13da-6ebf-41f2-abea-2688e9aefb5f",
8    amount      = 1200,
9    currency    = "INR",
10    reason      = "Purchase office supplies",
11    confidence  = 0.95,
12    merchant    = "Amazon Business",
13  )
14
15  if intent.status == "approved":
16    execute_payment(intent.id)        # ✅ proceed
17  elif intent.status == "pending_approval":
18    print("Awaiting finance team approval…")
19  else:
20    raise Exception(f"Denied: {intent.policy_decision}")

Regulatory Compliance

Built for the standards that matter.

Export audit trails in regulator-accepted formats — covering 8+ regulatory frameworks across major financial jurisdictions.

🇮🇳

India — RBI

Meets Reserve Bank of India requirements for AI-driven financial transaction audit trails and automated decisioning logs.

✓ Export-ready format

🇪🇺

EU — AI Act

Aligned with Article 13 transparency requirements for high-risk AI systems operating in EU financial services.

✓ Article 13 aligned

🇺🇸

USA — SOX

Sarbanes-Oxley export with control references, system-generated flags, and approval status for financial statement auditors.

✓ Section 302/404 ready

🇺🇸

USA — SOC 2

Immutable checksum-chained logs provide the evidence trail required for SOC 2 Type II security and availability audits.

✓ Type II evidence trail

🇬🇧

UK — FCA

Supports FCA Consumer Duty and Senior Managers & Certification Regime requirements for explainable AI-driven financial decisions.

✓ Explainability records

🇸🇬

Singapore — MAS

Aligned with the Monetary Authority of Singapore Technology Risk Management guidelines for AI systems in financial institutions.

✓ TRM aligned

🇦🇺

Australia — APRA

Meets APRA CPS 234 information security standards and APRA's AI risk management expectations for authorised deposit-taking institutions.

✓ CPS 234 aligned

🇦🇪

UAE — CBUAE

Supports Central Bank of UAE governance requirements for AI-assisted financial decisions, including audit trail and human oversight mandates.

✓ Governance-ready

View all compliance frameworks & country policies →

Simple Pricing

Pay for what you actually use.

One metric: spend intents processed per month. Start free, scale as your AI fleet grows.

Monthly Annual Save 20%

Starter

Free

Perfect for prototyping and small teams evaluating AI governance.

✓500 intents / month
✓3 registered AI agents
✓2 dashboard users
✓Policy engine
✓Audit trail
✓JSON export

Get Started Free