Plagiarism checker for marketing agencies.
Verify client deliverables before they leave your studio. Bulk-scan outsourced writers' submissions, AI-vet ghost-written content, integrate into your CMS pipeline, and export white-label PDF reports your client signs off on. Apache 2.0 engine — self-host if a client contract requires it.
Four agency roles. One QA pass before the client sees it.
Vet the deliverable before the client sees it
Catch issues before the Monday-morning client meeting.
- Bulk-check the week's deliverables in one queue
- White-label PDF report attached to delivery email
- Per-client AI-policy thresholds
Outsourced writer QA before paying
AI-spun content from contractors is the volume problem.
- Bulk scan a roster of freelancer submissions
- Per-writer history dashboard (catches repeat patterns)
- Citation-aware: properly-quoted brand voice doesn't inflate score
Self-check before delivering to the agency
The agency runs Noplag — run it first.
- Free tier covers a typical 2,500-word brief
- Self-citation: your own past pieces tag as SELF-CITE
- Match the agency's report shape before submission
Client-contract integrity policy
Some client contracts now require AI-disclosure.
- AI-policy threshold per client account
- Audit log of every deliverable check (exportable for compliance)
- Self-host option for SOC-2 / HIPAA / FedRAMP clients
Three steps before the deliverable email goes out.
Run on each draft, or batch-check a week's deliverables as a Monday QA pass. Either way the white-label PDF attaches to the delivery email as evidence of the originality check.
01 · Drafts come in
Freelancer / in-house / outsourced submissions land in your project-management tool (Asana, ClickUp, Trello, Notion, Linear). Configure the webhook to POST to /v1/checks on the deliverable-ready status transition. Audit log keeps the per-writer trail your editor-in-chief asks for monthly.
02 · Bulk scan + classify
The cascade scans against open web + Common Crawl + (optionally) your client's CMS back-catalogue. Citation classification separates legitimate quote-and-attribution from uncited paraphrase. Per-paragraph AI-likelihood flagging is coming soon — relevant because freelancer-AI submissions are now common enough that the per-writer pattern matters.
03 · White-label PDF + deliver
Export the report as a PDF with your agency's logo + brand colours + footer text. Attach to the delivery email as evidence of the originality check. Some clients now require this as part of the deliverable; for the others, it's a quiet differentiator that wins repeat contracts.
Six workflows agencies actually pay for.
From outsourced-writer QA to client-library hygiene — same engine, same report, configured per use case.
01 · OUTSOURCED WRITER QA
Vet contractor submissions before payment. Per-writer history shows repeat patterns: high AI scores from one freelancer over multiple jobs means a conversation. The audit log is what your editor-in-chief asks for at month-end.
02 · AI-SPUN CONTENT
Contractor-or-not, AI-generated content is a real volume problem. Per-paragraph AI-likelihood with ESL-adjusted band is coming soon — on the v1.2 roadmap. ESL-aware matters because international writers should not get unfairly flagged — Stanford 2023 documented the 61.3% FPR on naive detectors.
03 · CLIENT-LIBRARY AUDIT
Premium folder-level scan. Drop the client's published-content library into a folder; the engine builds a pairwise similarity matrix. Catches cross-page duplication (Helpful Content risk) and content recycled across multiple clients (something to surface to leadership).
04 · WHITE-LABEL PDF
Export the report with your agency's logo + brand colours + footer text. Per-client report templates (e.g., enterprise clients get a more detailed report than SMB clients). The PDF attaches to the delivery email as evidence; clients increasingly require this as part of their AI-disclosure policy.
05 · SEO CONTENT VETTING
Pre-publish for client blog posts: canonical-tag aware (your client's syndication network), Helpful-Content-Update-aware (catches the over-flagging-of-common-industry-terms that kills SEO content checkers). For SEO agencies producing 50+ articles/month per client, this is the workflow.
06 · GHOST-WRITE DETECTION
For ghostwritten content (memoirs, executive thought-leadership, books), the AI-likelihood signal (coming soon) is designed to help verify the ghostwriter actually wrote it rather than running it through an LLM. Sensitive use case — handle the conversation with the client carefully — but the signal is auditable.
Multi-client queue. Per-writer history. White-label-ready.
Group deliverables by client; sort by writer to spot patterns; bulk-approve the clean ones; escalate the rest. The audit log is per-writer, per-client, per-deliverable.
Your agency's logo on the PDF the client signs off on.
Logo + colour palette + footer
Upload your agency logo (SVG or PNG), set the primary + accent brand colours, configure the footer text (your agency address, contract reference, deliverable signoff line). Saved per-client so larger accounts get a more polished report than SMB ones.
Different report shapes for different contracts
Some clients want the full per-paragraph classification with source URLs. Others want a one-page executive summary. Pro+ supports per-client report templates — pick the right shape per account. The detection cascade is the same; only the layout differs.
Attach to email or embed in client portal
Export PDF, fetch a signed URL via API, or push directly into your client portal (HubSpot, Basecamp, ClientFlow). The signed URL expires after 7 days by default — the link your client gets in the delivery email won't be valid forever. Configurable per-client retention.
Two reasons agencies fold this into the contract — not as a cost line, as a quiet differentiator.
First — the white-label PDF attached to a deliverable email becomes a competitive moat. Clients increasingly require AI-disclosure or originality verification in their content contracts; agencies that quietly include it win the repeat business. Second — the per-writer history dashboard catches the AI-content-mill freelancer pattern before it ships under your agency's name. Both effects compound: the agency reputation rises (clients see the verification), the contractor pool gets cleaner (writers know they're being checked), and the AI-content workflow gets noticeably better than competitors' word-of-mouth quality. Per-deliverable cost on Premium at $79/mo for typical agency volume; Enterprise contracts negotiate down for 500+ deliverables/month.
Read the contract-language guideWhat agency leadership actually asks.
- Will our client deliverables be used to train your models?
- No. Drafts are fingerprinted (winnowing hashes, non-reversible signature) and stored in tenant-isolated database. The text itself is purged after 30 days unless retention is enabled per-client. Detection models are pre-trained on public corpora; we don't add client deliverables to any training set. Explicit in the DPA — your clients can require it if their contracts mandate it.
- How does the white-label PDF look on the client side?
- Your agency's logo on the cover, your brand colours throughout, your footer text. Zero Noplag visible attribution unless you explicitly opt in (some agencies do, for the procurement credibility). Per-client templates: Executive (1-page), Detailed (full report), Compliance (audit-trail only), AI-Disclosure (AI score + methodology). The PDF goes in the delivery email; the client sees your QA process, not your tooling vendor.
- Can we integrate with our project management tool?
- Pro+ tier — webhook integration with Asana, ClickUp, Trello, Notion, Linear. Configure the webhook to POST /v1/checks on the deliverable-ready status transition. The report attaches to the task as a metadata field. Sample integrations at github.com/NoplagLabs/noplag-engine/integrations.
- What about freelancer payment integration?
- Webhook + audit-log integration with the typical agency-finance stack: Bill.com, Tipalti, Wave, Xero. Configure a pre-payment hook that requires the originality check to pass before invoice approval. Stops the “paid the contractor for AI-generated content” failure mode before it costs you.
- Does per-writer history actually catch patterns?
- Yes — that's the workflow that wins repeat business. The dashboard tracks: average AI-score per writer over time, average uncited-paraphrase rate, frequency of overrides by editor, and the trend (is this writer's quality declining?). Catches contractors who started using AI mid-engagement. Catches contractors farming work to subcontractors who use AI. Catches the editor-of-record patterns that account managers want to know about.
- How does Noplag compare to Originality.ai for agency use?
- Originality.ai is the category leader for SEO-content marketing teams (their explicit positioning) — for that specific use case, they're the right tool. Noplag fits when your agency works in academic, multilingual, ESL, or self-hosted contexts where Originality's English-primary calibration and “not for students” disclaimer leave a gap. For pure SEO content marketing on English-only deliverables, both tools work; Noplag adds the open-source / self-host / per-language calibration.
- What's the per-writer rate limit on bulk scans?
- Pro tier handles up to 5,000 documents/month total across all writers. Enterprise scales to 25,000+/month with dedicated infrastructure. Per-writer there's no specific limit — your editorial-policy thresholds (e.g., flag a writer if their AI rate exceeds X% over 5 deliverables) are what drive the per-writer behaviour, not the platform.
- Can a client require us to use Noplag in our contract?
- Yes — agencies are starting to see this in client contracts (typically large clients with AI-disclosure or content-integrity requirements). Pair with a written AI-use policy from your agency and the report attached to delivery. Some clients require the report be exportable in compliance-grade audit format; that's the Compliance template at white-label PDF export.
- What about ghost-writing — sensitive use case?
- The AI-likelihood signal is coming soon (on the v1.2 roadmap) and is designed to help verify the ghostwriter actually wrote it (vs. running their draft through an LLM). Handle the conversation with the ghostwriter directly — the report is for your own QA, not for client distribution. Pro tier supports private folders that aren't exported to client reports.
- Can we self-host for clients with strict data-residency requirements?
- Apache 2.0 + Docker. Runs entirely inside your VPC. Useful for clients with SOC 2 / HIPAA / FedRAMP / EU-residency requirements where their content cannot go to a third-party cloud. The engine doesn't need internet access for indexed-corpus checks (the live-web verification layer is optional). Enterprise contracts include the corpus as an indexable bundle.
Try the queue. Bulk-check a week's deliverables.
Drop a folder of deliverables in. See the queue. White-label the PDF. Attach to your next delivery email. Free up to 2,500 words/check; Pro tier for the per-writer history dashboard, white-label PDF, and 500+ deliverables/month.