Stop AI agents
from burning money

One governance gateway that catches 11‑day loops, caps spend, and gives you Article 12‑ready traces for the EU AI Act

--- days

TO EU AI ACT HIGH‑RISK OBLIGATIONS

Article 12 & 14 record‑keeping, oversight and kill switches required

≈78%

ORGANISATIONS USE AI

Most have no loop‑aware governance or real‑time AI cost circuit breakers

Built to support EU AI Act‑aligned oversight and auditability for high‑risk AI systems

What Aqta Does

Runtime infrastructure for the age of adaptive governance

Adaptive Governance

Stops runaway agents in real-time. Detects infinite loops and halts execution before costs spiral or safety is compromised, reducing wasted LLM compute from redundant calls

EU AI Act Traces

Article 12‑ready logging. Immutable, tamper-evident audit trails that link every agent decision to a specific policy and user

Human-in-the-Loop

Meets Article 14 oversight requirements. Automatically routes high-risk decisions to human reviewers before execution

Spend Circuit Breakers

Hard caps on token usage per agent. Prevent "bill shock" with automatic cutoffs at 90% of budget and lower spend by avoiding unnecessary inference

Real-Time Visibility

Live observability into agent behavior. See exactly what your agents are doing, spending, and deciding right now

Zero-Code Integration

Drop-in gateway. Governance lives outside your application code, so you can switch frameworks without breaking compliance

How It Works

One‑line integration. No code changes. Works with OpenAI‑compatible SDKs and popular LLMs (Claude, Gemini, Llama) via the Aqta gateway

# Before

from openai import OpenAI

client = OpenAI(api_key="your-key")

# After (add Aqta governance)

client = OpenAI(
    api_key="your-key",
    base_url="https://proxy.aqta.ai"  # Aqta governance gateway
)

# All your existing agent code stays the same.

One‑line integration. No code changes to your agents. Aqta acts as an OpenAI‑compatible gateway that adds trace IDs, loop‑aware policies, cost and step circuit breakers and EU AI Act‑aligned audit logs

Every request flows through Aqta:

Your Agent → Aqta Gateway → OpenAI / Claude

Request arrives → Policy check → Risk score → LLM called → Response logged

↓ (if violation)

Human approval required / Agent halted.

If a request violates policy (PII, cost cap, toxic prompt), Aqta blocks it before it reaches the LLM and logs the decision for audit

Real-time monitoring dashboard shows:

✓ Agents active • Requests total • Loops detected • Cost breakdown

✓ Risk events • Decision audit trail • Spend vs. budget • Alerts

✓ Per-tenant isolation • Role-based access • Exportable compliance reports

Engineering, compliance and risk teams on the same page

Simple, SaaS Pricing

Predictable monthly plans with usage‑based overages for scale.

Starter

Free

For individual developers

Up to 10k requests/month
Basic loop detection
Cost tracking dashboard
7-day audit log retention
Community support

Built by Experts

Led by a founder with experience in integrity and governance programmes at leading consumer platforms and a background in GenAI research and full‑stack engineering

Integrity & governance at scale

Risk and policy experience in large consumer platforms

Full‑stack builder

Shipping the Aqta gateway and dashboard end‑to‑end

EU‑first

Based in Dublin, focused on EU AI Act readiness

Supported by advisers in EU AI Act & GDPR, cloud infrastructure, enterprise audit and AI safety research

Stop AI agents
from burning money