Sold by: Empirisch Tech GmbH
Deployed on AWS
Tells whether an English text was written by a person or by AI: a verdict per text (human, AI or mixed), the passages that read as AI, and a confidence level. Runs in your own AWS account on a real-time endpoint or batch transform, without network access.
Overview
ParaTrace tells whether an English text was written by a person or by AI. Deploy it in your own AWS account as a real-time endpoint or as batch transform jobs: SageMaker AI runs it without network access, nothing is stored, and your texts never leave your account.
What you get for every text
- A verdict: human, ai or mixed. A long text is labelled part by part, and the verdict is ai when at least 80% of it reads as AI, human when at least 90% reads as human, otherwise mixed.
- The passages that read as AI, as character offsets in the text you sent, and a score for each 512-token window.
- The probability that the text is AI-generated, and a confidence level (low, medium, high).
- Two settings: strict (about 1 in 100 human texts flagged, the default) and lenient (about 5 in 100, and fewer AI texts missed).
Use cases
- Education: check essays and assignments on a learning platform or in a grading workflow.
- Publishing and media: screen submissions, articles and reviews.
- Trust and safety: find AI-written reviews, posts and profiles at scale with batch transform.
- Compliance and HR: check applications, reports and statements where authorship matters.
How to use it
- Real-time: JSON requests with up to 10 texts of up to 200,000 characters each; plain text works too.
- Batch transform: JSON Lines files in Amazon S3, one text per line, answered line by line.
- Recommended instance type: ml.g4dn.xlarge (one NVIDIA T4) for both. Deploy it with SageMaker AI's newer image: InferenceAmiVersion al2023-ami-sagemaker-inference-gpu-4-1 for an endpoint, TransformAmiVersion al2-ami-sagemaker-batch-gpu-535 for batch transform. The sample notebook does it for you.
Training data More than 2 million texts: human-written texts from many kinds of writing (essays, abstracts, sections of scientific papers, books, short fiction, blog posts, emails, speeches, recipes, patents, open textbooks), each paired with an AI text on the same topic written by one of more than 20 language models; human passages paired with AI paraphrases of them; answers of chat assistants; and AI texts rewritten by "humanizer" prompts.
Accuracy, on texts held out from training
- 1.3% of 2,081 human texts from ten kinds of writing flagged at strict; 1.9% of the UChicago benchmark's human texts.
- 99.6% of 26,086 answers from more than 100 chat models caught; 4.4% of the Epoch AI benchmark's AI texts missed, all of them imitations of a human author's style.
- A text judged ai with high confidence was human 0.1% of the time; judged human with high confidence, it was human 99.3% of the time (texts half human, half AI).
Speed on ml.g4dn.xlarge (NVIDIA T4)
- One text of 331 words in 0.14 s; 40 texts with 10,607 words in one request in 2.95 s: about 3,600 words a second.
- An endpoint starts in about 5 minutes.
Limitations
- English only. Verdicts are estimates, not proof of who wrote a text.
- Texts under 50 words: unreliable (the answer warns).
- Human persuasive essays are the weak spot: 18.4% flagged at strict.
- Text rewritten by "humanizer" tools is often missed: 36% of text rewritten by StealthGPT is caught at strict.
- AI passages spread into the human text next to them, so ai_fraction runs high on mixed texts.
Highlights
- Private by design: deployed in your own AWS account, run without network access, nothing stored. Your texts never leave your account.
- Verdicts you can act on: human, AI or mixed for every text, with the AI-written passages marked, a probability and a confidence level.
- Fast on one NVIDIA T4: about 3,600 words a second on ml.g4dn.xlarge, in real time or in batch transform jobs.
Details
Sold by
Delivery method
Latest version
Deployed on AWS
New
Introducing multi-product solutions
You can now purchase comprehensive solutions tailored to use cases and industries.
Features and programs
Financing for AWS Marketplace purchases
AWS Marketplace now accepts line of credit payments through the PNC Vendor Finance program. This program is available to select AWS customers in the US, excluding NV, NC, ND, TN, & VT.
Pricing
Pricing is based on actual usage, with charges varying according to how much you consume. Subscriptions have no end date and may be canceled any time.
Additional AWS infrastructure costs may apply. Use the AWS Pricing Calculator to estimate your infrastructure costs.
Dimension | Description | Cost/host/hour |
|---|---|---|
ml.g4dn.xlarge Inference (Batch) Recommended | Model inference on the ml.g4dn.xlarge instance type, batch mode | $1.00 |
ml.g4dn.xlarge Inference (Real-Time) Recommended | Model inference on the ml.g4dn.xlarge instance type, real-time mode | $1.00 |
Vendor refund policy
No refunds: ParaTrace is charged by the hour it runs in your account, and hours used are not refunded.
Legal
Vendor terms and conditions
Upon subscribing to this product, you must acknowledge and agree to the terms and conditions outlined in the vendor's End User License Agreement (EULA) .
Content disclaimer
Vendors are responsible for their product descriptions and other product content. AWS does not warrant that vendors' product descriptions or other product content are accurate, complete, reliable, current, or error-free.
Delivery details
Amazon SageMaker model
An Amazon SageMaker model package is a pre-trained machine learning model ready to use without additional training. Use the model package to create a model on Amazon SageMaker for real-time inference or batch processing. Amazon SageMaker is a fully managed platform for building, training, and deploying machine learning models at scale.
Deploy the model on Amazon SageMaker AI using the following options:
Version release notes
First release on Amazon SageMaker AI: real-time endpoints and batch transform on ml.g4dn.xlarge (NVIDIA T4), in JSON, JSON Lines and plain text.
Additional details
Inputs
- Summary
Input summary English texts as JSON (real-time: up to 10 texts of up to 200,000 characters each), as JSON Lines (batch transform: one text per line) or as plain text (one text).
Input description application/json: {"text": "...", "texts": ["...", ...], "operating_point": "strict" or "lenient", "mode": "detailed" or "fast"}. operating_point strict (the default) flags about 1 in 100 human texts, lenient about 5 in 100. mode detailed (the default) adds each text's windows and segments; fast answers the verdicts only. application/jsonlines (batch transform, SplitType Line): one text per line, as a JSON string or as {"text": "...", "document": "a name", "operating_point": ..., "mode": ...}. text/plain: one text.
- Input MIME type
- application/json, application/jsonlines, text/plain
{
"texts": [
"Four score and seven years ago our fathers brought forth on this continent, a new nation, conceived in Liberty, and dedicated to the proposition that all men are created equal.\n\nNow we are engaged in a great civil war, testing whether that nation, or any nation so conceived and so dedicated, can long endure. We are met on a great battle-field of that war. We have come to dedicate a portion of that field, as a final resting place for those who here gave their lives that that nation might live. It is altogether fitting and proper that we should do this.\n\nBut, in a larger sense, we can not dedicate -- we can not consecrate -- we can not hallow -- this ground. The brave men, living and dead, who struggled here, have consecrated it, far above our poor power to add or detract. The world will little note, nor long remember what we say here, but it can never forget what they did here. It is for us the living, rather, to be dedicated here to the unfinished work which they who fought here have thus far so nobly advanced. It is rather for us to be here dedicated to the great task remaining before us -- that from these honored dead we take increased devotion to that cause for which they gave the last full measure of devotion -- that we here highly resolve that these dead shall not have died in vain -- that this nation, under God, shall have a new birth of freedom -- and that government of the people, by the people, for the people, shall not perish from the earth.",
"Remote work has fundamentally reshaped how organizations approach productivity and collaboration. By eliminating daily commutes, employees can reclaim valuable time and often report a better work-life balance. At the same time, companies gain access to a broader talent pool that is no longer constrained by geography.\n\nHowever, remote work also presents unique challenges. Without the spontaneous interactions of a shared office, teams may struggle to maintain a strong sense of connection and culture. Clear communication becomes essential, and leaders must be intentional about fostering trust and inclusion.\n\nTo make remote work successful, organizations should consider the following strategies:\n\n- Establish clear expectations around availability and response times.\n- Invest in reliable collaboration tools and digital infrastructure.\n- Schedule regular check-ins to support both performance and well-being.\n- Create opportunities for informal social connection.\n\nUltimately, the most effective approach is often a thoughtful hybrid model that combines the flexibility of remote work with the benefits of in-person collaboration. By prioritizing communication, trust, and employee well-being, organizations can build resilient teams that thrive in any environment."
],
"operating_point": "strict"
}
{"document": "gettysburg-address", "text": "Four score and seven years ago our fathers brought forth on this continent, a new nation, conceived in Liberty, and dedicated to the proposition that all men are created equal.\n\nNow we are engaged in a great civil war, testing whether that nation, or any nation so conceived and so dedicated, can long endure. We are met on a great battle-field of that war. We have come to dedicate a portion of that field, as a final resting place for those who here gave their lives that that nation might live. It is altogether fitting and proper that we should do this.\n\nBut, in a larger sense, we can not dedicate -- we can not consecrate -- we can not hallow -- this ground. The brave men, living and dead, who struggled here, have consecrated it, far above our poor power to add or detract. The world will little note, nor long remember what we say here, but it can never forget what they did here. It is for us the living, rather, to be dedicated here to the unfinished work which they who fought here have thus far so nobly advanced. It is rather for us to be here dedicated to the great task remaining before us -- that from these honored dead we take increased devotion to that cause for which they gave the last full measure of devotion -- that we here highly resolve that these dead shall not have died in vain -- that this nation, under God, shall have a new birth of freedom -- and that government of the people, by the people, for the people, shall not perish from the earth."}
{"document": "remote-work-answer", "text": "Remote work has fundamentally reshaped how organizations approach productivity and collaboration. By eliminating daily commutes, employees can reclaim valuable time and often report a better work-life balance. At the same time, companies gain access to a broader talent pool that is no longer constrained by geography.\n\nHowever, remote work also presents unique challenges. Without the spontaneous interactions of a shared office, teams may struggle to maintain a strong sense of connection and culture. Clear communication becomes essential, and leaders must be intentional about fostering trust and inclusion.\n\nTo make remote work successful, organizations should consider the following strategies:\n\n- Establish clear expectations around availability and response times.\n- Invest in reliable collaboration tools and digital infrastructure.\n- Schedule regular check-ins to support both performance and well-being.\n- Create opportunities for informal social connection.\n\nUltimately, the most effective approach is often a thoughtful hybrid model that combines the flexibility of remote work with the benefits of in-person collaboration. By prioritizing communication, trust, and employee well-being, organizations can build resilient teams that thrive in any environment.", "operating_point": "lenient"}
{"document": "pride-and-prejudice", "text": "It is a truth universally acknowledged, that a single man in possession of a good fortune, must be in want of a wife.\n\nHowever little known the feelings or views of such a man may be on his first entering a neighbourhood, this truth is so well fixed in the minds of the surrounding families, that he is considered the rightful property of some one or other of their daughters.\n\n\"My dear Mr. Bennet,\" said his lady to him one day, \"have you heard that Netherfield Park is let at last?\"\n\nMr. Bennet replied that he had not.\n\n\"But it is,\" returned she; \"for Mrs. Long has just been here, and she told me all about it.\"\n\nMr. Bennet made no answer.\n\n\"Do you not want to know who has taken it?\" cried his wife impatiently.\n\n\"You want to tell me, and I have no objection to hearing it.\"\n\nThis was invitation enough.", "mode": "fast"}
"Photosynthesis is the remarkable process by which plants, algae, and certain bacteria convert light energy into chemical energy. During this process, chlorophyll in the chloroplasts absorbs sunlight, which powers the transformation of carbon dioxide and water into glucose and oxygen. This not only provides the energy that fuels plant growth but also releases the oxygen that sustains most life on Earth. In essence, photosynthesis serves as the foundation of nearly every food chain, making it one of the most important biological processes on our planet."
Resources
Vendor resources
Support
Vendor support
AWS infrastructure support
AWS Support is a one-on-one, fast-response support channel that is staffed 24x7x365 with experienced and technical support engineers. The service helps customers of all sizes and technical abilities to successfully utilize the products and features provided by Amazon Web Services.
Similar products
AI-powered help desk and ticketing system that unifies email, chat, messaging, and SMS in one inbox for support teams of any size.
Text - Customer Service Platform with AI Agent, Live Chat, Help Desk built to capture leads, support customers and increase online sales.
Detect the dominant emotion in English feedback, tickets and social posts. Runs inside your AWS account, so no text is sent to a third-party API. 93.2% accuracy across six emotions on 4,000 held-out examples.
The Mphasis AI for Text Processing service enables enterprises to generate actionable insights from Unstructured text content from a variety of sources such as annual reports, customer reviews, documents, forms, emails, tweets, chat transcripts, news articles etc. We leverage our patented 'AI for Text Processing' solution to engage with clients through assessments, implementations and managed services offerings. We help enterprises target impactful use cases to improve so that the derived insights can drive business benefits. Our Assessment, Workshops and Implementations focus on specific parts of the most relevant use cases and help define the exact benefits to be derived from the initiative such as efficiency, cost, customer experience etc.
Customer reviews
No customer reviews yet
Be the first to review this product . We've partnered with PeerSpot to gather customer feedback. You can share your experience by writing or recording a review, or scheduling a call with a PeerSpot analyst.