Product
How it actually works
A deeper look at compliance runs, rule sets, rendering, OCR, and the evidence that backs every verdict.
The core insight
HTML is not what recipients see
A standard HTML audit passes when the right text exists somewhere in the source. But required disclaimers, logos, and prominence relationships only matter when they're visible in the rendered creative — and that often means inside images.
RuleRun renders the email, extracts visible content using computer vision, and evaluates your policy against what a recipient would actually see.
Core concepts
Rule Sets
A Rule Set is a named collection of compliance rules that maps to a policy or regulatory requirement. You define required assets, forbidden text, font-size relationships, approved domains, and content guidelines. One Rule Set can be reused across every campaign that falls under the same policy.
Compliance Runs
A Compliance Run is a single evaluation of one email or creative against a Rule Set. Each run produces a verdict (pass/fail/warning), a per-rule breakdown, and a PDF audit report with annotated evidence. Runs are triggered by email, REST API, or automation workflows.
Render and OCR
The email is rendered as a recipient would see it — not as raw HTML. Computer vision extracts visible text from the rendered output, including text embedded inside promotional banners and images that HTML inspection cannot reach.
Evidence and reports
Every rule check captures evidence: the exact measured value, the configured threshold, and an annotated screenshot marking the relevant region. The PDF report compiles this evidence for human review and audit trail.
API and automation
All functionality is available via REST API. Trigger runs from your CI/CD pipeline, campaign approval tool, or AI agent workflow. Webhooks and polling are available to receive results programmatically.
The output
A compliance report you can act on
Every run produces a PDF report showing the overall verdict, individual rule outcomes, measured values, policy thresholds, and annotated screenshots.
Reviewers see exactly what was checked and what evidence supports the verdict — enough to approve, reject, or escalate without running the check manually.