Writing and content decision

GPTZero

Don't rebuild: GPTZero's primary defensible value is its proprietary detection model and institutional trust/partnerships, which a lone developer or small team cannot realistically reproduce; a narrower open-source detector UI is possible but will lack the claimed accuracy, integrations, and production polish.

Visit website
SubscriptionCustom pricing
Initial build30 hours
Monthly upkeep5 hours + $400
Evidence2/3 runs agree

No open-source build does this yet

Nothing published replaces this one, so a replacement starts from an empty file. Here is what it would have to cover.

What a replacement has to do

  • Accept pasted text or uploaded document, run an AI-detection model to compute per-document and per-sentence AI-likelihood scores, highlight suspected AI phrases and generate a downloadable report.

What it still won’t have

  • Proprietary detection model and claimed industry-leading accuracy
  • Built-in integrations and extensions (Chrome extension, Google Docs, Canvas, Zapier)
  • Human verification / expert feedback workflows and writing-replay video features
  • Scale, reliability, and dataset advantages of a production vendor

What remains hard

  • Proprietary modelsGPTZero uses its own proprietary model that takes hundreds of factors into consideration and boasts the highest detection accuracy in the industry
  • Brand trustOfficial AI detector partners of the American Federation of Teachers
Read the build prompt

First-year cost

No published price

GPTZero does not publish a price we could read, so there is nothing to compare against. What building costs is below.

Money you would actually spend

Keep paying

Subscription price × seats × 12

Build it

AI build APIs + hosting

Time you would spend

What you would spend

What we assumed

The verdict above measures whether you could build it. This one is only about money.

Runnable build prompt

Not run yet
Build a minimal AI-detection web service using Node.js (Express) for the API, a React UI, and Postgres for storage. In scope: (1) POST /scan that accepts text or file, (2) integrate an open-source detection model (use one of the candidate GitHub projects) to compute per-sentence and document-level AI-likelihood scores, (3) store scan metadata and results in Postgres, (4) web UI to paste text, show per-sentence highlights and a download-as-PDF report, (5) simple API GET /scans/:id to retrieve results. Out of scope: Chrome extension, Google Docs integration, human-review workflows, training new detection models. Require input validation, error handling for file uploads and model failures, and unit tests for the API and integration tests for the end-to-end scan flow.
How we checked3 sources · 2/3 runs agreed · evidence score 20

How the score was reached

  • Pay verdict base20
  • 3 cited sources+3
  • Hard moats found in the evidence-3
  • Evidence score20

The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time — so the same evidence always produces the same number.

How scoring works →

Cited sources · 3

Every page the run actually retrieved.

Integrity checks

What held up, and what did not.

! 2 of 3 runs agreed; the verdict is the majority✓ Citations limited to fetched pages! 2 moats quoted from the page