Requires manual text input
The tool does not fetch URLs. Paste the full policy text manually. Incomplete input produces incomplete analysis.
GDPR and PDPA core only
Does not cover ePrivacy Directive, AI Act, NIS2, or national implementation laws (e.g. BDSG, Italian Privacy Code). Gaps under these frameworks are not detected.
Area-level verdicts, not clause-level
Each verdict covers an entire compliance area. The excerpt field helps locate relevant text, but does not replace a clause-by-clause legal review.
English-language policies only
Prompts and reasoning are in English. Analysis of policies in other languages may produce degraded results. No warning is shown when this occurs.
AI output is not legal advice
High confidence does not mean legal certainty. The model can misread ambiguous clauses or miss context that a qualified lawyer would catch.
EDPB guidelines not auto-updated
Prompts reflect GDPR interpretation as of May 2026. New enforcement decisions or EDPB guidelines are incorporated manually — not in real time.
Very long policies may degrade
Extremely long policies (with multiple addenda) are sent to the model in a single call. Analysis quality may decrease for content in the latter sections.
API key required for AI mode
AI analysis requires an Anthropic API key. Without it, the tool falls back to keyword matching (v1.0 engine), which cannot detect semantic gaps.
This tool is a student research project. It is designed to support — not replace — professional compliance review. It does not constitute legal advice and is not a substitute for a qualified DPO or privacy lawyer. For professional use, treat results as a structured starting point for further analysis.