EU AI Act Article 5 and Article 50. TrustLint offline regex: not the hosted policy engine, not a system classifier. Check already-produced text (model input or output) against ComplyEdge TrustLint offline YAML rules (regex corpus). Use for post-hoc evaluation of text that already exists; for pre-generation screening of a prompt about to be sent, use scan_prompt. Argument text must be non-empty content to evaluate (empty/whitespace is rejected before any rule runs; this is not a system-risk questionnaire). Optional jurisdiction is EU, US, or GLOBAL; omit it to evaluate the full corpus. GLOBAL is a separate corpus tag, not the union of EU and US : reuse the same token you used with list_rules for a scoped check. Runs fully offline : no network call, no API key, no credentials stored, no writes outside the process, read-only and idempotent; regex match against the bundled corpus; does not classify AI-system risk tiers or call any remote API. Returns PASS/FAIL with rule ID, severity, citation, and remediation.

EU AI Act Article 5 and Article 50. TrustLint offline regex: not the hosted policy engine, not a system classifier. List TrustLint offline compliance rules in the ComplyEdge corpus : the discovery tool; returns no PASS/FAIL or SAFE/RISK verdict. Use it to scope a jurisdiction before calling check_compliance; do not use it to evaluate text. Optional jurisdiction is EU, US, or GLOBAL; omit it to list the full corpus. GLOBAL is a separate corpus tag, not EU∪US : pass that exact token into check_compliance next for a consistent scoped evaluation. Returned total is the post-filter count. Runs fully offline : no network call, no API key, no credentials stored, no writes outside the process, read-only and idempotent. Returns rule IDs, titles, severities, jurisdictions, and categories.

EU AI Act Article 5 and Article 50. TrustLint offline regex: not the hosted policy engine, not a system classifier. Scan an AI prompt with ComplyEdge TrustLint offline YAML rules (regex corpus) before generation : pre-generation only. Use when the prompt is about to be sent; for post-hoc evaluation of already-produced text, use check_compliance. Argument prompt must be a non-empty candidate prompt about to be sent to a model (empty/whitespace is rejected; not already-produced output). This tool has no jurisdiction filter : it always evaluates the full bundled corpus. Runs fully offline : no network call, no API key, no credentials stored, no writes outside the process, read-only and idempotent; regex match against the bundled corpus; does not classify AI-system risk tiers or call any remote API. Returns SAFE or RISK_DETECTED with cited findings from the offline regex corpus.