Pangram

What it's good for

• 1

  Academic-integrity reviews where a false accusation against a student is unacceptable

• 2

  Hiring teams checking written application materials without wrongly flagging real candidates

• 3

  Newsrooms and publishers verifying contributor copy is human-written

• 4

  Compliance and provenance checks in regulated content workflows

• 5

  High-stakes disputes where the detector's finding has to hold up to scrutiny

Pricing

• Individual

  For solo and light use

  $20/mo

• Pro

  3,000 credits

  $65/mo

• API

  Pay-as-you-go, about $0.05 per scan

  Usage-based

• Institutional

  Volume and integration for organizations

  Custom

How to use it

Paste or upload text into the web app for a single check, or send it through the API for batch and programmatic use at about $0.05 per scan. Read the result as a probability, not a verdict, and lean on Pangram's very low false-positive rate when a call has real consequences. For institutions, the custom tier covers volume and integration into existing review pipelines. As with any detector, pair it with human judgment: use it to flag work for a closer look, not to hand down an automatic decision.

Pros & cons

Frequently asked questions

• How much does Pangram cost?

  Individual plans run about $20/mo, Pro is around $65/mo for 3,000 credits, and the API is pay-as-you-go at roughly $0.05 per scan. Institutional pricing is custom for volume and integrations. There is no confirmed free tier, so plan on a paid plan from the start.

• Pangram vs GPTZero: what is the difference?

  Both detect AI-written text, but they aim at different priorities. GPTZero is the accessible, freemium entry point that also covers images and humanized content. Pangram's differentiator is defensible accuracy: independent validation from University of Chicago Booth and University of Maryland, and a very low false-positive rate of about 1 in 10,000. When a wrong accusation carries real consequences, Pangram is built for that; for broad, casual checks GPTZero is easier to start with.

• Is Pangram free?

  No. There is no confirmed free tier. Plans start at about $20/mo for individuals, with Pro around $65/mo and API access at roughly $0.05 per scan. If you need a free option for ad-hoc checks, a freemium detector like GPTZero is the place to look, though it trades away some of Pangram's accuracy.

• What is Pangram best for?

  High-stakes AI-text detection where a false accusation is the real risk: academic integrity, hiring, publishing, and compliance. Its independent academic validation and roughly 1-in-10,000 false-positive rate make its findings easier to defend than a general-purpose detector's.

• How accurate is Pangram?

  Its standout claim is a very low false-positive rate, about 1 in 10,000, and that claim has been independently validated by researchers at University of Chicago Booth and University of Maryland. No detector is perfect, so treat results as strong probabilities and keep a human in the loop, but the low false-positive rate is what sets it apart from general-purpose tools.

More

• Pangram home
• Pangram API docs

Alternatives to Pangram

Other tools we'd consider for the same job.

• GPTZero

  AI text + deepfake detection. Covers writing, images, and humanised AI content for compliance, education, and trust workflows

• Originality.AI

  AI detection + plagiarism scanning for content marketing and SEO. Lower false positives on long-form than general-purpose detectors

Other AI Detection & Trust tools

• Copyleaks

  Enterprise, multilingual AI-content and plagiarism detection with an API, built for education and compliance teams

• GPTZero

  AI text + deepfake detection. Covers writing, images, and humanised AI content for compliance, education, and trust workflows

• Originality.AI

  AI detection + plagiarism scanning for content marketing and SEO. Lower false positives on long-form than general-purpose detectors