Security & trust
The security story is not a PDF of promises — it is enforced by the pipeline, on every job.
The rules the pipeline enforces
- Pixels never come from AI. Rendering is deterministic, tested code applying your committed brand specification. AI may author content where you allow it; it never draws your deliverable.
- Every artifact carries its egress class.
zero-egressorcontract-protected— written by the pipeline based on what actually happened to the job, never by hand. A job that touched an AI API cannot be labelled zero-egress. - The zero-egress tier contains no AI at all. No model, no cloud, no internet in the pipeline — verified at installation, on your box, offline.
- AI authoring is fail-closed. On the intelligence tier, if the DPA attestation is not in place, AI calls refuse to run. Spend is capped per job.
- Synthetic media is consent-gated and disclosed — always. Video featuring a presenter likeness or a cloned voice requires recorded consent before the pipeline will produce it, and the deliverable discloses that it is synthetic. No product flag, tier, or entitlement can bypass this.
- Nothing unverified is delivered. The visual-QA gate runs on every job, in every tier. A failing deliverable is withheld, with findings.
Provenance you can audit
Every deliverable records exactly which version of your brand specification produced it. When your brand changes — new logo, new palette, new template — you can prove which documents were produced under which rules, months later.
What “verified” means (and doesn’t)
Verified means the deliverable passed the visual-QA gate: layout, branding, logo placement, text fit. It does not mean DocMark fact-checked the content — your authors (human or AI) own the substance, and web-research outputs ship with cited sources so your reviewers can check them.
Want the detail? The security one-pager covers the trust boundaries, the egress classes, and the consent gates in a form your security team can file.