From draft to scan

1. Paste or upload the draft

Drop in a landing page, brochure, or campaign block. PDF (text layer), DOCX, or plain text. Set language and market hints when you know them; otherwise leave Auto.

Prefer 100+ words. Shorter text can still scan, but estimates are noisier — Lockmark warns you instead of faking certainty.

2. Choose the client policy

Advisory only — report the estimate and let your team decide.
Grammar-assist OK — allow light assist tools; still flag generative patterns.
No generative copy — tighter ship-gate when the contract forbids generative drafting.

3. Read the scan

You get a band (likely human / uncertain / mixed / likely machine), an AI-likeness score, hotspots on a text map, forensic evidence, and suggested next steps. Optional brand voice check and revision comparison against approved pages or the previous draft.

4. Human in the loop

Open flagged excerpts in the draft, rewrite, or export a scan report for the client file. Lockmark is not designed for unattended mass detection or legal determinations.

5. Chrome extension (included)

Highlight copy on a live CMS, Docs, or client page and scan without pasting into Studio. Same AI scan and plagiarism check, same free-tier character limits. After a scan, use Open in Studio for policy modes, hotspots, and reports.

1. Open chrome://extensions → enable Developer mode.
2. Click Load unpacked and select the web/chrome-extension folder from your Lockmark install (pilots get this path from us).
3. Set the extension’s API / Studio base URL to your Lockmark host (local default http://127.0.0.1:8001).
4. On a normal webpage, select text → Scan selected text.

Public Chrome Web Store listing comes later; pilots use unpacked or an unlisted build.

What the score is built from

A hybrid of deep language signals (mDeBERTa when enabled) and writing-style / forensic cues, with guards for short text and certain agency landing-page registers in Dutch and German. Raw model scores stay available under “How this score was built.”