Analysis Credits

Buy Credits, Run
Premium Analyses

Power your RL analysis with credits. 1 credit = 1 analysis unit. Buy a bundle, use them whenever you need — no subscription, no monthly charge, credits never expire.

Starter pack: $19.90 for 105,000 credits  ·  Volume bundles save up to 100% more

View Credit Bundles
Credits never expire
Secure payment via Stripe
One-time payment, no subscription
What you get

Built for teams shipping real AI

Every Premium feature is designed to solve the alignment problems that actually break RL systems.

Automatic Parameter Adjustment

Detects reward imbalances and automatically updates reward weights mid-run. No manual tuning, no wasted epochs.

Live Training Monitoring

Real-time dashboard showing reward dynamics, hacking signals, and alignment scores as your model trains — step by step.

Alignment Guardrails

Active enforcement prevents reward hacking and misalignment from propagating. Think of it as a seatbelt for your training loop.

Advanced Anomaly Detection

Multi-layer anomaly detection using statistical and ML-based methods. Catches subtle drift that simple thresholds miss entirely.

Priority Support

Direct email and chat support from our alignment engineers. Average first-response time under 2 hours during business hours.

Compare plans

Free vs Premium

See exactly what you get when you upgrade.

Feature
Free
Premium
Core Analysis
Reward distribution analysis
Imbalance & drift detection
Warnings & recommendations
Alignment reports (PDF export)
Training steps analyzed
100k
Unlimited
Premium — Auto-Alignment
Automatic reward rebalancing
Live monitoring during training
Continuous alignment enforcement
Advanced anomaly detection
Priority support (< 2hr response)
Custom amount

Need a specific amount?

Enter how many credits you need. Base rate: 100,000 credits = $19.90 ($0.000199/credit).

Minimum 5,000 credits ($1.00). Credits never expire.

════════════ -->
Choose your bundle

Buy Analysis Credits

Credits power every premium analysis run. Buy once, use anytime — no subscription, no expiry. Bigger bundles give you more credits per dollar.

Starter
+5% Credits
$19.90

one-time · no subscription

105,000 credits
+5% extra credits — 5,000 bonus credits
  • 210× Small-scale RL agent analyses
  • 52× basic Atari agent analyses
  • Great for trying premium features
Popular
$49.90

one-time · no subscription

275,000 credits
10% off — 25,000 bonus credits
  • 550× Small-scale RL agent analyses
  • 137× basic Atari agent analyses
  • 27× robotics control policy analyses
Best Value
$99.90

one-time · no subscription

715,000 credits
30% off — 215,000 bonus credits
  • 1,430× Small-scale RL agent analyses
  • 357× basic Atari agent analyses
  • 71× robotics control policy analyses
  • 14× large-scale multi-agent analyses
100% Bonus
$299.90

one-time · no subscription

1,500,000 credits
100% bonus — 750,000 extra credits
  • 3,000× Small-scale RL agent analyses
  • 750× basic Atari agent analyses
  • 150× robotics control policy analyses
  • 30× large-scale multi-agent analyses

Analysis reference

Small-scale RL agent
≈ 500 credits per analysis
Basic Atari agent
≈ 2,000 credits per analysis
Robotics control policy
≈ 10,000 credits per analysis
Large-scale multi-agent
≈ 50,000 credits per analysis

Checking your balance…

Just getting started?

The Starter Pack

Get 105,000 credits instantly for just $19.90 — the lowest-cost way to try premium analyses, with 5% extra credits included. No subscription, no recurring charge. Perfect for running your first analysis or evaluating premium features.

  • 105,000 credits deposited instantly after payment (5% bonus included)
  • Use them to run premium training sessions with full auto-alignment
  • One-time payment — no monthly charge, no subscription required
  • Credits never expire and are tracked in your dashboard
  • Great for evaluating premium features before committing
One-time purchase
$19.90

one-time · no subscription

105,000 credits included
+5% Extra Credits Included

Secure payment via Stripe  ·  Instant delivery

FAQ

Common questions

Do credits expire?
No. Credits never expire. Once purchased, they stay in your account indefinitely and are consumed only when you run a premium analysis. You can buy a bundle now and use the credits months later.
How does the auto-adjustment feature work?
When RewardGuard detects a reward imbalance or hacking pattern, it calculates corrected reward weights using a combination of statistical analysis and learned priors. These corrections are applied to your training loop at configurable intervals (default: every 500 steps). You can set boundaries on how aggressively it adjusts, or run in suggestion-only mode first.
What RL frameworks does RewardGuard support?
RewardGuard has native integrations for Stable-Baselines3, RLlib, CleanRL, and custom PyTorch/JAX loops. The core library uses a framework-agnostic callback interface, so if your framework supports callbacks (most do), you can integrate RewardGuard in under 10 lines of code.
Is my training data sent to your servers?
No. RewardGuard runs entirely in your environment as a Python library. Your raw training data, model weights, and environment state never leave your infrastructure. Only license validation (a single authenticated request at startup) communicates with our servers.
Do you offer team or enterprise plans?
Yes. For teams of 5 or more, we offer volume discounts and consolidated billing. For larger organizations that need SSO, audit logs, on-premise deployment, or custom SLAs, please email us at support@rewardguard.dev and we'll set up a call.
What's included in the license key after purchase?
Your Premium license key unlocks the rewardguard[premium] package, which includes the PremiumMonitor, auto-adjustment engine, live monitoring dashboard, and all advanced detection algorithms. The key is tied to your account and can be used across all your machines.