Pangram

What it does

Pangram's AI Detection API takes a block of text and returns a classification: overall fractions of text that look AI-written, AI-assisted, or human-written, plus per-section windows with labels, scores, and character ranges. Developers call it through a Python SDK, using predict() for single texts and submit_bulk() for asynchronous batches, or through the REST API.

The Pangram MCP server exposes check_for_ai, check_for_similarity, and list_models tools to compatible agent clients, authenticated through Pangram account sign-in rather than a copied API key. Typical uses include moderation queues, LMS submissions, CMS review flows, and backfills of content archives. Results can also include a dashboard link for sharing.

Access requires a Pangram account and prepaid API credits from $5. Realtime checks are limited to 5 QPS, and bulk jobs receive a 20% discount for large document sets. Enterprise contact is the route for SOC 2, zero data retention options, and higher rate limits.

What it can do

• Detect AI-generated and AI-assisted text with overall fraction scores
• Return per-window labels, confidence, and character ranges for analyzed sections
• Process single texts in realtime and large sets via asynchronous bulk jobs
• Generate shareable dashboard links for results
• Check text for similarity with online sources through the MCP server
• Expose detection tools to compatible agent clients via OAuth-secured MCP

Works with

• Google Docs

Before you choose

• Model availability depends on the API key; new code should pass an explicit model rather than relying on the default, since the argument becomes required after September 30, 2026
• Realtime rate limit is 5 QPS; bulk jobs suit large document sets and run asynchronously
• MCP usage draws from the same API credit balance as the API, not the dashboard scan balance

Listed under

• Data
• AI detection
• API
• Education
• Image detection
• Plagiarism checker
• Publishers

Get started

Create an account, generate a key in the API tab, and purchase prepaid credits.

1. Install the Python SDK with pip install pangram-sdk
2. Set PANGRAM_API_KEY as an environment variable or pass the key to the Pangram constructor
3. Call list_models() and pass a returned model name, such as pangram-4, to predict() or submit_bulk()

Developer API uses prepaid credits purchasable from $5 to $2,000. The realtime rate for Pangram 4 is $0.05 per 100 words, and bulk jobs receive a 20% discount. The Individual and Professional plans on the pricing page also list monthly API credits ($10 and $200 respectively).

Source: pangram.com/solutions/api, checked 29 Oct 2026. Prices can change; check before you buy.

Sources for this overview

Product pages, documentation, and pricing checked 29 Oct 2026. Each source supports part of the overview; availability and prices can change.

• AI Detection API | Pangram ↗www.pangram.com
• Pangram MCP Server | Pangram ↗www.pangram.com
• Python Quickstart - Pangram Labs ↗docs.pangram.com
• AI Image Detector - Check Images for AI Generation | Pangram ↗www.pangram.com
• Pricing | Pangram ↗www.pangram.com

Company website

pangram.com: product, pricing and policy pages

YC directory mirror

Batch and company name

Social accounts

Accounts the company links, and its posts on X

Human review

Checked by a person

knowledge_base

manual_review

Last checked 29 Oct 2026. Every field keeps its source and date; unknown means we couldn’t confirm it.