Skip to main content
Intentra
BetaGet 3 Months Free of Intentra Pro for joining the Beta. Book your spot →

Get Refunds for AI Coding Mistakes

Intentra detects when Cursor, Claude, Gemini, and other AI coding tools do things you didn't ask for — then helps you get your money back.

$brew install atbabers/intentra/intentra
100 free scans/monthNo credit card required

From detection to refund in minutes

See how Intentra finds refund opportunities

terminal
$ curl -fsSL https://install.intentra.sh | sh
✓ Intentra installed successfully
$ intentra install cursor
✓ Hooks installed for Cursor
# ... work normally in Cursor ...
$ intentra scan today
Today: 15 scans ($4.50)
3 anomalies detected ($1.20 discrepancy)
Potential savings: 27%

Zero-friction setup

One command to install. One command to enable tracking. Start seeing your AI costs immediately.

Instant visibility

Every scan tracked in real-time. See exactly what you're spending and where anomalies occur.

Actionable insights

Anomalies highlighted automatically. Know immediately when tool behavior diverges from expectations.

What exactly is a "scan"?

A scan is the LLM calls that occur between prompts. Even if the AI retries multiple times, you only pay for 1 scan.

"Fix this bug"
= 1 scan
"Write tests"
= 1 scan
"Refactor code"
= 1 scan

💡Why scans, not tokens?

No hidden costs: Even if the AI retries 5 times, you pay for 1 scan
Easy to understand: "I made 15 requests today" vs "I used 47,329 tokens"
Predictable billing: Know exactly what you'll pay before you hit upgrade

Technical note: A scan is 3-5 LLM API calls1 grouped by user intent. We handle the complexity of hook aggregation, so you only pay for meaningful actions.

1 Average based on analysis of 10,000+ real AI coding sessions across Cursor, Claude Code, GitHub Copilot, Antigravity, Windsurf, and Gemini CLI. Simple tasks (typo fixes) typically use 1-2 calls, complex refactors use 5-8 calls. We use the median range for billing simplicity.

Built For Developers

Catch unauthorized AI behavior, build evidence, and get refunds from your coding tools

Privacy First

We never store your prompts, code, or AI responses. Data is encrypted in transit, processed for anomalies, then immediately discarded.

20 Detection Patterns

Catch ignored instructions, retry loops, excessive thinking, and 17 more patterns that indicate your AI tool did something you didn't authorize.

Cost Intelligence

Identify overcharges and billing anomalies. Export structured evidence to build your refund case with providers.

Real-Time Dashboard

Track your AI coding costs as they happen. See exactly where money was wasted and which scans qualify for refunds.

Simple, transparent pricing

Unlimited scans for paid plans. No surprises.

Free

$0/mo

Perfect for trying out Intentra

  • 100 scans per month
  • 3 devices
  • AI-powered detection*
  • Cross-scan pattern analysis
  • Export evidence reports
  • Basic support
Popular

Pro

$20/mo

For professional developers

  • Everything in Free, plus:
  • Unlimited scans
  • AI-powered detection
  • Automated dispute remediation(Coming Soon)
  • Priority support

Enterprise

Custom

For teams and organizations

  • Everything in Pro, plus:
  • Analytics
  • Audit log streaming
  • SSO & RBAC
  • Dedicated support

* AI-powered detection on Free plan limited to first 25 scans

Trusted by developers

See what early users are saying about Intentra

Intentra identified $47 worth of retry loops in our first week across our engineering org.

SC
Sarah Chen
Senior Engineer, TechStartup*

Finally, visibility into AI coding costs. The anomaly detection alone paid for itself.

MR
Michael Rodriguez
Principal Engineer, DataCorp*

As a bootstrapped start-up, we're constantly using AI tools to build quickly. Intentra makes sure we get all the value, without any wasted spending.

JC
John Collins
Founder, Yogen
Verified User

Used Intentra to identify $48 in billing discrepancies from retry loops and excessive tool behavior in the month of February 2026.

Verified User
Founder, Intentra ✓

* Testimonials from private beta participants.

Join 50+ developers1 tracking their AI coding costs

$0+2
Discrepancies identified
0+3
Scans tracked
0%4
Avg cost reduction

1 Early adopters from private beta (December 2025 - March 2026). Includes free tier users.

2 Cumulative discrepancies identified across all users during private beta.

3 Total scans processed across all users since private beta launch in December 2025.

4 Average cost reduction from early adopters who addressed detected anomalies. Individual results may vary based on tool usage patterns and anomaly frequency.

Stop Paying for AI Code You Didn't Ask For

Install in 30 seconds. Detect billing anomalies. Get refunds.

$brew install atbabers/intentra/intentra
100 free scans/month
No credit card required
Cancel anytime
Intentra