Appraysal
Compliance Intelligence Engine

Demystifying the
Approval Process

Our mission is to help you ship faster and reduce rejection risk. We use continuously refreshed validation datasets and verified rejection evidence to turn a black-box process into a measurable pre-submission check.

Raw Code
Validation
Approval
01

The Foundation

We codified Apple and Google policy rules that are practical for static analysis, from manifests and permissions to metadata and policy-sensitive code patterns, including Apple 1.2 user-generated content controls.

Large Repo Corpus
02

The Engine

We run evidence-first detection and compare outputs against validation sets and verified rejection examples to tune confidence, severity, and rule behavior.

Historical Analysis
03

Continuous Calibration

Rules are regularly updated as policies change. Each rule is assigned a precision tier (Definite, Likely, or Advisory) based on detection confidence, helping you focus on what matters most.

Tiered System

By the Numbers

Our scanner is calibrated using real-world data, not theoretical guidelines.

ContinuouslyUpdated
Compliance Rules
iOS & Android
345
Verified Rejections
With source evidence
652
Published Apps
With store URLs
Evidence-Calibrated
Findings
Validated on dual 500-repo evaluation sets

Precision-Based Tiers

Not all detections are equal. We assign each rule a confidence tier based on detection precision.

Accuracy rankings treat support=0 FP-only rules as a separate pathology list, not part of normal worst-rule F1 ordering.

Definite
High-confidence evidence-backed

Configuration checks and file existence rules. If triggered, this is strongly associated with rejection risk and should be fixed first.

Likely
Moderate-confidence multi-signal

Multi-signal pattern detection. High probability of rejection based on pattern correlation across code.

Advisory
Variable precision

Broader patterns and heuristics. Flagged for manual review—false positives are possible.

New Capability

The Vision Engine

Reviewers don't just read code; they look at your app. We now use multimodal AI to "see" what they see, catching visual compliance issues before submission.

Context-Aware Scanning

Our engine doesn't just look for "Lorem Ipsum". It understands context. It knows that a "Login" screen is fine, but a screenshot showing only a login screen (Guideline 2.3.3) is a rejection.

Cross-Verification

We correlate code presence with visual evidence. Only here can we detect if your code contains In-App Purchases (Guideline 3.1.2) but your screenshots fail to show a compliant paywall.

Metadata Intelligence

Metadata Validation

Most rejections aren't caused by code bugs, but by "unforced errors" in App Store Connect. We automatically parse your Fastlane configuration to catch these before you even open a browser.

Privacy & Lexical Checks

We verify your Privacy Policy URL's reachability and scan your app description for accidentally pasted legal jargon (EULA) that triggers Guideline 2.3.1.

Screenshot Completeness

Our engine checks if you've included all required device sizes for both iPhone and iPad, preventing the "Missing Screenshots" roadblock.

Submission Readiness

Automatic validation of review notes, demo account credentials, and release note quality to improve submission readiness before manual review.

Radical Transparency

No scanner is perfect. We believe you deserve to know exactly where our blind spots are. While we provide industry-leading coverage for code, privacy, and metadata issues, we cannot replicate the human elements of App Review.

Static Code AnalysisHigh Confidence

We catch many API/config issues with high signal quality; critical paths still require manual verification.

Privacy & EntitlementsHigh Confidence

Deep inspection of manifests and capabilities.

Runtime BehaviorNo Coverage

We cannot see dynamic content or server responses.

Subjective QualityNo Coverage

We can't judge design aesthetics or app value.

Static CodePrivacyMetadataSubjectiveRuntime