Content Detector AI provides online AI-generated text detection services to help educators and content operators quickly identify ChatGPT, Claude, Gemini and other outputs.

ContentDetectorAI

Core parameters and statistics of Content Detector AI

Content Detector AI initially provided AI text detection services as an independent website at contentdetector.ai, and was later migrated to the Moxby Agent market to be delivered in the form of a plug-in. The current version is 1.4.6 (released on 2026-05-05, last updated on 2026-07-16). The core positioning of the product is an "AI writing style review assistant" rather than an "AI author identifier" - it identifies mechanical writing characteristics by analyzing the perplexity, burstiness and model fingerprint of the text, and presents the results in a sentence-by-sentence color score.

Projects	Public Information
Official positioning	Editorial aid for robotic writing pattern review
Tool type	AI writing quality review (content detection + humanized rewriting)
Delivery form	Moxby Agent plug-in (formerly independent web tool)
Current version	1.4.6 (released on 2026-05-05, last modified on 2026-07-16)
Detection method	Perplexity score + Burstiness mode + Model fingerprint
Analysis granularity	Sentence-by-sentence color confidence scoring
Input requirements	Recommend more than 100 words
Number of installations	27 (Moxby market open)
Developer	Abigail Foster (Moxby Pro author)
Price	Free
Place of Belonging	United States

Product form change: Content Detector AI was originally an independent online detection site (contentdetector.ai), and currently the domain name has been 301 redirected to the Moxby market page. The product is transformed from an independent web tool into a built-in plug-in for the Moxby browser Agent, relying on the Moxby runtime to run contextually. This means users will need to install the Moxby desktop client (supported for Mac and Windows) before using the detector.

Detection mechanism: Similar to mainstream AI detectors (such as Originality.ai, GPTZero), Content Detector AI makes probabilistic inferences based on statistical features. The core difference is that it does not claim to be a "deterministic judgment", but is positioned as a "writing style review aid". Among the three detection dimensions, perplexity measures the unexpectedness of the text (AI-generated text is generally less perplexing), burstiness measures the variability in sentence length and structure (human writing is more bursty), and model fingerprint attempts to match the output distribution characteristics of a specific LLM.

Accuracy Limitation: The official statement clearly states that it does not guarantee the accuracy of the test. The results of short texts (<100 words) have limited reference value; for texts that have been deeply rewritten, dense with professional terms, written in non-native languages, or created by a mixture of humans and machines, the detection signal may be fuzzy or unreliable.

User and market recognition of Content Detector AI

The current market data for Content Detector AI is relatively early, and publicly verifiable information mainly comes from the Moxby market page.

Installation Size: The Moxby Market page publicly displays a total of 27 installations, with no user reviews yet. This size shows that the product is in the early stage of promotion and has not yet formed large-scale word-of-mouth communication.

Developer Background: The author Abigail Foster is marked as a Pro author on the Moxby platform, and her profile mentions "focusing on AI detection, plagiarism analysis, and content quality assurance workflows." This account has only one listing on the Moxby market and has not yet formed a tool matrix.

Ecological dependence: The availability of the product is directly limited by the user base of the Moxby platform itself. As a relatively new browser agent platform, Moxby's user base is much smaller than that of Chrome extensions or independent SaaS products. Therefore, the actual reach of Content Detector AI is constrained by the ceiling of the platform's installed capacity.

Comparison with similar tools:

Comparison Dimensions	Content Detector AI	Originality.ai	GPTZero
Delivery form	Moxby plug-in (requires client)	Standalone Web SaaS	Standalone Web + Chrome extension
Pricing	Free	Paid subscription (starting at $14.95/month)	Free credits + paid plans
Positioning	Writing style review assistance	Professional AI content detection	Educational academic integrity detection
Detection transparency	No accuracy claims	Claims 99%+ accuracy	Claims 99%+ accuracy
Batch detection	Unpublished (Moxby workflow)	Support batch API	Support batch upload
Overwrite/Humanize	Built-in AI Humanizer	None	None

Market Positioning Judgment: Content Detector AI is currently closer to a "content quality auxiliary tool" than a "serious AI detection tool". Its differentiation lies in the fact that it does not exaggerate the detection accuracy and has built-in humanized rewriting capabilities. It is suitable for quality self-inspection scenarios in the editing process, but it is not suitable as a source of evidence for academic sanctions or legal disputes.

Cost Advantages of Content Detector AI

Content Detector AI is currently completely free, but the cost advantage needs to be evaluated in the context of the Moxby platform - there is no subscription fee for the tool itself, but use requires installing the Moxby desktop client and accepting the context in which it runs.

C client/individual users: The tool itself has zero cost and no hidden subscription levels. Individual writers, editors, and content creators can install and use the full range of detection and user-friendly rewriting features directly from the Moxby Marketplace for free. The cost is mainly reflected in the installation and learning cost of the Moxby client (about 5-10 minutes), as well as the consumption of underlying LLM resources called by Moxby Agent during the detection process (depending on the resource allocation strategy of the Moxby platform, whether the frequency is limited or limited is subject to Moxby's official policy).

Developer/API integration: Content Detector AI does not provide independent API endpoints, and the detection capabilities are encapsulated in the Moxby Agent plug-in and cannot be called independently. This means that developers cannot embed it into their own pipelines or do batch automation integration. If the team needs API-level AI detection capabilities, it needs to evaluate solutions such as Originality.ai (providing REST API, starting at $14.95/month) or GPTZero (providing API, billed as per volume).

Enterprise/Institution: The current version does not provide enterprise-level deployment or privatization options. The Moxby platform's own support strategy for enterprises (SSO, audit logs, data residency) is subject to Moxby's official public information. Content Detector AI's offline detection feature (analysis is done in a local Moxby Agent session) has some appeal for enterprise data privacy, but it lacks independent SLA and enterprise-level management capabilities.

Real Cost Deduction: For an editorial team that detects an average of 50 short articles per day, the explicit cost of using Content Detector AI is zero, but the implicit costs include IT support time for all team members to install the Moxby client, manual review time due to uncertainty in detection results (since the official does not make accuracy commitments, key decisions still require manual secondary verification), and the time-consuming article-by-article operations caused by the lack of batch processing capabilities. In contrast, although paid competitors have a single monthly subscription fee, the batch API and higher detection confidence may dilute the overall cost of a single detection in the actual workflow.

Main functions of Content Detector AI

Content Detector AI integrates AI detection and writing quality improvement into the same Moxby Agent plug-in, and the function link is "Scan → Position → Rewrite Suggestions".

• Real-time AI Scan: Start the detection directly in the Moxby Doc editor, analyze the text sentence by sentence, and mark sentences that appear to be mechanical writing. Detection does not rely on external servers, all analysis is done within the local Moxby Agent session, and the text never leaves the user's device. Perfect for instant self-checking during the writing process, without having to switch tabs or copy-paste to a third-party page.

• Sentence-by-Sentence Confidence Score: The detection results are color-coded to mark the probability of AI writing features of each sentence, with green indicating "natural", yellow indicating "needs attention", and red indicating "highly mechanical". Users can visually locate high-risk passages instead of just getting an overall percentage. This granularity is valuable in the editing process—it allows you to pinpoint the parts that need to be rewritten, rather than having to reinvent the wheel.

• AI Humanizer Rewriting Guide: In addition to detection, the built-in AI Humanizer skill can guide the Agent to naturally rewrite marked paragraphs. The direction of rewriting includes adding specific cases, improving the diversity of sentence patterns, strengthening transitional cohesion, and removing general filler words. The goal is not to "fool the detector" but to improve the readability and realism of the content. This feature puts "problem discovery" and "problem resolution" in the same workflow, shortening the cycle from detection to modification.

• Local Privacy Handling: Detection analysis is done locally through the Moxby gateway, and user text is not stored or logged. For scenarios with high content confidentiality requirements (such as unpublished business manuscripts and internal reports), this design reduces the risk of data leakage and is a key difference from most cloud detection tools.

• Multi-model source compatibility: Text from any source such as ChatGPT, Claude, Gemini, etc. can be input, including purely human-written samples. The tool itself does not do "origin tracing" - it does not distinguish which model the text comes from, it only evaluates whether the writing style contains mechanical characteristics. This means it works equally well with rewritten or remixed text, but it also requires more human judgment to interpret the results.

Functional synergy: AI Scan locates problematic paragraphs → Colored annotations present priorities → AI Humanizer provides rewriting direction, and three sections are closed in the Moxby Doc editor. Users do not need to repeatedly switch between detection tools and editing tools, nor do they need to judge "what to do after detection" by themselves. This "detection is the starting point for improvement" design is closer to the actual editing workflow than a tool that simply gives a percentage score.

Content Detector AI model and version evolution

The version information of Content Detector AI is disclosed through the Moxby market page, and the product has evolved from an independent website form to a Moxby Agent plug-in form.

Current version: 1.4.6 (released on 2026-05-05, last modified on 2026-07-16). The market page shows that this version contains 1 Tool (AI Scan) and 0 standalone Skills (AI Humanizer is released as a built-in capability rather than a standalone Skill). Version number iteration can reflect plug-in function updates, but the official changelog has not been published in full.

Version node context:

Version Node	Date	Major Changes
Independent website stage	~2024-01	contentdetector.ai operates as an independent online detection tool, providing basic AI text judgment
Initial version of Moxby plug-in	2026-05-05	Release v1.0, migrated to Moxby market, and redelivered in the form of Agent plug-in
The latest version v1.4.6	2026-07-16 (last modified)	Continuous iteration, stable functions, no major architectural changes

Version evolution characteristics: The transformation from an independent website to a platform plug-in means a fundamental change in the product architecture - the detection logic of the original independent version may be processed by its own backend, but after migration, it will be run locally by Moxby Agent. This change improves privacy protection capabilities, but also ties the product’s availability to the life cycle of the Moxby platform. The version numbers from 1.0 to 1.4.6 indicate that the product is still in the feature iteration stage, but there is no public roadmap.

Iteration Suggestion: Before official adoption, it is recommended to pay attention to the last modification date and version number changes on the Moxby market page. If there are no updates for more than 3 months, it may mean that developers have reduced their maintenance efforts and need to evaluate long-term availability risks.

Technical advantages of Content Detector AI

Content Detector AI's technical route is designed around "lightweight localization detection + editing process closure", and the core mechanism can be broken down into three levels.

Detection mechanism: Perplexity + Burstiness + Fingerprint joint analysis

The detection engine uses a combination of three signals: Perplexity score measures the statistical surprise of the text - AI models tend to select high-probability tokens when generating, so the overall perplexity of AI text is usually lower than human writing; Burstiness measures the variability of sentence length and sentence structure - human writing exhibits higher rhythm fluctuations, while AI output tends to have more uniform sentence structure; model fingerprint attempts to match the output distribution pattern of a specific LLM (such as ChatGPT's paragraph structure preference vs Claude's word selection preference). The combined use of the three is more robust than a single dimension, but the official statement does not guarantee accuracy.

Effects and Applicable Scenarios: This detection mechanism has the best recognition effect on "pure AI output without modification". All three signals are less distinguishable for writing that has undergone extensive manual rewriting, is dense with technical terminology, or is written by a non-native English speaker. This is also the technical reason why the official proactively stated that it "does not claim to accurately determine the author" - it is not a lack of product capabilities, but a theoretical uncertainty boundary in the statistical testing itself.

Local processing architecture: Detection analysis is done locally on the user through the Moxby gateway, and the text does not leave the device. This means that detection capabilities in context without network connectivity may be limited (depending on the local reasoning capabilities of Moxby Agent), but privacy and security are better than cloud detection solutions. For scenarios where you are dealing with unpublished manuscripts, internal reports, or confidential materials, local infrastructure is a significant compliance advantage.

Technical comparison with competing products:

Technical Dimension	Content Detector AI	Originality.ai	GPTZero
Detection kernel	Perplexity + Burstiness + fingerprint	Multi-model ensemble classifier	Perplexity + Burstiness
Processing location	Local (Moxby Agent session)	Cloud API	Cloud API
Accuracy Statement	No Claim	Claimed 99%+	Claimed 99%+
Rewriting assistance	Built-in AI Humanizer	None	None
Detection granularity	Sentence-by-sentence color scoring	Overall + sentence-by-sentence	Overall + sentence-by-sentence
Minimum valid input	~100 words	~50 words	~100 words

Engineering Features: As a Moxby plug-in, the detection capabilities of Content Detector AI are exposed to Moxby Agent in the form of a Tool. Large models can call the Tool through the Agent workflow to complete detection. This architecture allows detection capabilities to be orchestrated into more complex automated processes - for example "detect articles → mark high-risk paragraphs → automatically trigger Humanizer rewriting → output a report of modification suggestions". However, due to the multi-Agent orchestration capabilities of the Moxby platform, whether the current plug-in supports this chain call is subject to the official actual function.

How to use Content Detector AI

Content Detector AI currently has the only entry point: through the Moxby desktop client. The original independent website contentdetector.ai has been redirected to the Moxby marketplace page.

Installation and usage steps:

1. Download Moxby client: Visit Moxby official website to download the desktop client (supports Mac and Windows) and complete the installation.
2. Install the AI ​​Content Detector plug-in: Enter the Agent Marketplace in the Moxby client, search for "AI Content Detector", and click Install (free).
3. Open Doc Editor: Create or open a document in Moxby, and select the text content that needs to be detected.
4. Execute AI Scan: Trigger detection through the "AI Scan" button or Agent command in the editor, and the plug-in will analyze it sentence by sentence and mark the confidence color.
5. View results and rewrite: Locate high-risk sentences based on colored annotations, use the AI ​​Humanizer capability to get rewriting suggestions, and adjust the text as needed.

Typical workflow duration deduction:

Task steps	Manual completion	Using Content Detector AI	Efficiency changes
Manual quality inspection of an 800-word article	10-15 minutes (read through + judge based on experience)	2-3 minutes (scan + view annotations + rewrite key paragraphs)	shortened by about 75%
Check mechanical expressions sentence by sentence	Based on intuition, possible omissions	Automatic annotation sentence by sentence, no omissions	Improved coverage
The direction of subsequent modifications is determined	You need to judge the modification points by yourself	There are AI Humanizer suggestions as a starting point	Modification efficiency is improved

Usage Precautions:

• Test results are for editorial reference and should not be used as final evidence of AI identification.
• The ideal input length is more than 100 words, and the annotation confidence of short text may not be reliable enough.
• Detection is network independent (local processing), but AI Humanizer overrides may require Moxby Agent to call LLM, which may require a network connection.
• Since the product no longer has an independent web version, it cannot be used without installing the Moxby desktop version.

Product Pricing for Content Detector AI

Content Detector AI is completely free in its current form, with no tiered subscriptions or hidden paywalls.

Pricing Panorama:

User level	Fees	Function coverage	Restrictions
C-side/personal writer	Free	AI Scan detection + AI Humanizer rewriting	Moxby client needs to be installed; detection frequency is subject to Moxby platform policy
Content team/editing team	Free	Same as above (multiple users each install the client)	No team management function; no batch processing panel
Enterprise/Institutional	Free (currently)	Same as above	No enterprise SLA; No SSO/Audit; No private deployment

The truth about the free model: The main reason why the product is free is that it exists as a plug-in in the Moxby platform ecosystem, and its commercialization is carried by the Moxby platform itself. Developer revenue may come from the platform sharing mechanism of Moxby Pro authors, rather than charging end users. This model is user-friendly, but the long-term dependence on the Moxby platform’s incentive policy for plug-in developers and the sustainability of the platform itself.

Hidden Costs:

• Moxby client dependency: Moxby desktop must be installed before it can be used. Moxby's own resource usage (memory CPU) needs to be included in the cost.
• No API/SDK: Cannot be integrated into your own content management or publishing pipeline, batch scenarios need to be operated article by article.
• Uncertainty of results: Since the official does not guarantee accuracy, key decision-making scenarios still require manual review. This time cost may exceed the time saved by the test itself.

Application scenarios of Content Detector AI

The core application scenarios of Content Detector AI focus on "content quality self-inspection" rather than "AI author identification". The following three types of scenarios are best suited to maximize its value:

• Content editing and pre-publication quality review: Before the manuscript is published, the editor performs a mechanical scanning of the writing characteristics of the AI-assisted generated content to locate paragraphs with repetitive sentence patterns, abrupt transitions, and lack of specificity, and obtains rewriting directions through AI Humanizer. Key points of verification: Submit pure human writing samples and AI-assisted samples together, and observe whether the annotation differences of the detector are consistent with the editor's own judgment, so as to calibrate the trust level.

• SEO and Marketing Content Optimization: Before publishing AI-assisted blogs, landing pages, and promotional copy, SEO teams and content marketers use detectors to scan for “AI cavities”—paragraphs that are too general and lack specific data and case support. Humanizer's rewriting guidelines help transform content into a usable manuscript that is more compliant with EEAT standards. Verification focus: Test on different topics (technical tutorials vs brand stories vs product descriptions) to confirm whether the sensitivity of the detector to various styles is consistent.

• Quality standardization of multi-author content platforms: When content platforms or operating agencies face manuscripts submitted by multiple writers, they need a unified mechanical writing inspection scale. The results provided by Detector can be used as a reference signal in the internal quality review process, and cooperate with the final judgment of human editors to help the team maintain the unity of content style. Key points of verification: Establish a comparison relationship between internal "AI feature scoring" and manual quality scoring, and find out the threshold range suitable for your own content type.

Unsuitable Scenarios:

• Irreversible decisions such as academic punishment/recruitment elimination - the official statement clearly states that the results should not be used as the only basis, and any disciplinary decision based on test results carries the risk of misjudgment.
• High-frequency detection of short texts (<100 words) - the results of social media short copywriting, comments, titles and other scenarios have limited reference value.
• Automated Pipelines that require API batch integration - no independent API endpoints and cannot be embedded into existing content management processes.
• Scenarios with stringent compliance requirements for detection accuracy (e.g. legal documents, financial disclosures) – current tools’ accuracy claims cannot meet these needs.

Applicable groups of Content Detector AI

Content Detector AI’s free positioning and editing assistance determines that it is most suitable for the following four types of roles:

• Independent writers and content creators: Creators who need to produce blog newsletters and long social media articles on a daily basis can use it to perform mechanical self-checks on writing before publishing. Prerequisite: Familiar with the basic operations of the Moxby client, and willing to transfer part of the writing process to the Moxby Doc environment. Does not fit the boundary: If the writing process has been fixed in tools such as Notion and Google Docs and is unwilling to be migrated, each inspection requires copying and pasting, and the workflow friction is high.

• Content Editors and Proofreaders: Editors responsible for reviewing multiple manuscripts and needing to quickly locate homogeneous expressions. Sentence-by-sentence color annotation helps editors prioritize passages most likely to deviate from natural expression. Prerequisite: Have basic editing judgment and be able to distinguish the priority of "detection annotation" and "manual judgment" - annotation is only for reference, and the final modification decision should still be based on editing experience.

• SEO & Marketing Operations: Operations role responsible for maintaining brand content quality and EEAT compliance. A round of mechanical writing testing can be done before the AI-assisted content goes online, and targeted optimization can be done based on Humanizer recommendations. Prerequisite: Understand the requirements for "originality, specificity, and author expertise" in the EEAT standards, and be able to judge whether Humanizer recommendations are consistent with the brand voice.

• Educators (auxiliary use): Teachers can use the detector as a teaching aid when discussing writing style to show "what kind of sentences are easily judged as mechanical writing". Important restriction: Test results cannot be used as evidence of academic misconduct - the official statement has explicitly excluded this use, and the statistical nature of the tool determines that it cannot reliably identify a single piece of work.

Not applicable to the crowd:

• Legal and Compliance Review Team: In scenarios that require high-certainty, auditable AI content determination, the accuracy statement and localized delivery form of the current tool do not meet compliance requirements.
• Enterprise Procurement Decision Maker: If you have scaled deployment API integration, audit logs, and SLA requirements that the current version of Content Detector AI cannot meet, you should evaluate Originality.ai or GPTZero Enterprise.
• Non-English content workers: The product interface and detection model are mainly optimized for English texts. The detection results in Chinese or other languages ​​are subject to official actual performance, and no public explanation has been made.

Human-machine collaboration boundary:

Sectional	Degree of automation	Description
Text scanning and annotation	100% automation	Agent automatically completes sentence-by-sentence analysis without manual intervention
Result interpretation and prioritization	Semi-automatic	Color marking provides a basis for sorting, but whether high-risk paragraphs really have problems requires manual judgment
Determination of rewriting direction	Semi-automated	AI Humanizer provides rewriting suggestions, but the final modification plan requires manual confirmation
Publish/delivery decisions	Human labor is irreplaceable	The final decision on whether content meets publishing standards must be made by the editor or person in charge

Summary and Outlook of Content Detector AI

Content Detector AI has taken a differentiated route among AI detection tools: it does not pursue the digital competition of "identification accuracy", but positions itself as a "writing quality review assistant in the editing process" and integrates detection and humanized rewriting into the same workflow. Free pricing, local privacy processing, and sentence-by-sentence visual annotation are its three main attractions.

Current Limitations and Uncertainties:

• The product has been transformed from a standalone web tool to a Moxby plug-in, and its usability is completely dependent on the Moxby platform. If the Moxby platform strategy adjusts, ceases operations, or modifies plug-in policies, the tool will be directly ineffective.
• There is only one developer, and the 27 installations indicate that the product is in a very early stage, and the long-term maintenance capabilities and update rhythm have not been fully verified.
• The detection model lacks independent third-party benchmark evaluation. All accuracy statements are based on the description on the official page. Users cannot compare their performance horizontally on different test sets.
• No API, no batch processing, and no team management functions, which limits its application in large-scale scenarios.

Summary of positioning differences with competing products:

Selection Dimensions	Content Detector AI	Originality.ai	GPTZero
Best adaptation scenario	Personal editing quality self-examination	Professional content detection + API integration	Educational academic integrity screening
The biggest advantage	Free + local privacy + built-in rewriting	High accuracy + enterprise-level functions	Reputation in the education industry + free quota
Biggest shortcoming	Extremely early stage + platform binding	Paid and not cheap	Limited free quota
Purchase Threshold	Zero	Monthly Subscription	Free/Subscription
Applicable organization size	Individual to small team	Medium to large team to enterprise	Educational institution to enterprise

Procurement/Adoption Risk Assessment: For individual writers and editors, the threshold for zero-cost trial is extremely low. It is recommended to test the consistency of the test results with manual judgment on 1-2 typical manuscripts before making a decision on whether to incorporate it into the daily process. For content teams or organizations, it is recommended to do the following verifications before formal adoption: (1) Do a double-blind test with 20-30 sample manuscripts from known sources (including three types of pure manual, pure AI, and hybrid rewriting), and statistically detect the label distribution of the three categories; (2) Evaluate the team deployment cost and compatibility of the Moxby client (whether it runs normally in the existing IT environment); (3) Pay attention to the update frequency and developer activity of the plug-in in the Moxby market page - if it continues If there are no version updates for more than 3 months, long-term usability should be re-evaluated. Core recommendation: Do not use it as the only AI content judgment tool. It is more suitable as a reference signal in the editing quality process, complementing manual review.

Related tools: , gptzero

Version Info

• Content Detector AI Online ：There is no official precise date yet. Continuously updated online testing services. （~2026-07）
• Content Detector AI Beta ：There is no official precise date yet. The early version provides basic AI content determination. （~2024-01）