AI Content Detector - GPT, Claude, Gemini Text Check
Pricing
Pay per usage
AI Content Detector - GPT, Claude, Gemini Text Check
Detect AI-generated text using statistical analysis. Analyzes perplexity, burstiness, sentence patterns, vocabulary diversity, and more. No external API key needed.
Pricing
Pay per usage
Rating
0.0
(0)
Developer
George Kioko
Actor stats
0
Bookmarked
4
Total users
1
Monthly active users
14 days ago
Last modified
Share
AI Content Detector - Detect ChatGPT, Claude, Gemini & AI-Generated Text
Detect AI-generated text using advanced statistical analysis. This AI text detector analyzes writing patterns, sentence structure, vocabulary diversity, and linguistic markers to determine if content was written by a human or generated by AI models like ChatGPT, Claude, Gemini, GPT-4, Llama, or other large language models. No external API key needed — runs entirely on statistical analysis.
Features
- No API key required — Pure statistical analysis, no external dependencies
- Instant results — Analyzes text in milliseconds via Standby mode HTTP API
- 9 detection signals — Sentence uniformity, burstiness, vocabulary diversity, AI phrases, sentence starters, readability, paragraph structure, repetitive patterns, punctuation
- Confidence scoring — 0-100 AI score with high/medium/low confidence ratings
- Verdict classification —
ai_generated,likely_ai,mixed,likely_human, orhuman_written - Bulk analysis — Analyze up to 20 texts in a single request
- Detailed breakdown — See exactly which signals triggered and why
- Works for all AI models — Detects patterns common to ChatGPT, Claude, Gemini, GPT-4, Llama, Mistral, and more
- Real-time API — Use as a fast HTTP API endpoint via Apify Standby mode
How It Works
This AI content detector uses 9 statistical signals to identify AI-generated text:
| Signal | What It Measures | Why AI Differs |
|---|---|---|
| Sentence Uniformity | Variation in sentence lengths | AI produces uniform lengths |
| Burstiness | Dramatic changes between sentences | Humans are more "bursty" |
| Vocabulary Diversity | Type-token ratio | AI hits a predictable middle range |
| AI Phrase Density | Known AI transition phrases | AI overuses "furthermore", "delve", "crucial" |
| Sentence Starters | How sentences begin | AI uses predictable starters |
| Readability | Flesch-Kincaid grade level | AI targets grade 10-14 consistently |
| Paragraph Structure | Paragraph length consistency | AI creates uniform paragraphs |
| Repetitive Patterns | Structural repetition | AI repeats syntactic patterns |
| Punctuation Patterns | Comma usage patterns | AI uses commas very consistently |
Input
| Field | Type | Description |
|---|---|---|
text | String | Single text to analyze (min 30 characters) |
texts | Array | Multiple texts to analyze in bulk (max 20) |
Input Example
{"text": "The rapid advancement of artificial intelligence has fundamentally transformed the way we approach complex problem-solving across various domains."}
Bulk Input Example
{"texts": ["AI has transformed how we work and live in the modern world.","lol i cant believe the game last night was so crazy my friend was screaming"]}
Output
{"text": "The rapid advancement of artificial intelligence has fundamentally...","ai_score": 82,"human_score": 18,"verdict": "ai_generated","confidence": "high","word_count": 145,"sentence_count": 6,"details": {"scores": {"sentence_uniformity": { "score": 78, "weight": 15 },"burstiness": { "score": 85, "weight": 12 },"vocabulary_diversity": { "score": 72, "weight": 10, "ttr": 0.621 },"ai_phrases": { "score": 90, "weight": 20, "ai_phrases_found": 8, "examples": ["furthermore", "crucial role", "various domains"] },"sentence_starters": { "score": 65, "weight": 10 },"readability": { "score": 88, "weight": 8, "grade_level": 12.3 },"paragraph_structure": { "score": 70, "weight": 8 },"repetitive_patterns": { "score": 55, "weight": 10 },"punctuation_pattern": { "score": 80, "weight": 7 }},"text_stats": {"avg_sentence_length": 24.2,"sentence_length_std_dev": 3.8,"vocabulary_diversity_ttr": 0.621,"avg_syllables_per_word": 1.82,"flesch_kincaid_grade": 12.3,"commas_per_sentence": 2.1,"paragraph_count": 2}}}
Verdict Scale
| AI Score | Verdict | Meaning |
|---|---|---|
| 80-100 | ai_generated | Very likely written by AI |
| 60-79 | likely_ai | Probably AI with some human elements |
| 40-59 | mixed | Could be either, or AI-assisted human writing |
| 20-39 | likely_human | Probably human-written |
| 0-19 | human_written | Very likely written by a human |
Use Cases
- Academic integrity — Check student submissions for AI-generated content
- Content verification — Verify that blog posts, articles, and reports are human-written
- SEO compliance — Detect AI content that may be penalized by search engines
- Hiring — Verify cover letters and writing samples are authentic
- Publishing — Screen submissions for AI-generated manuscripts
- Social media — Detect AI-generated comments and bot accounts
- Journalism — Verify sources and submissions are genuinely human-written
API Endpoints (Standby Mode)
Use the actor in Standby mode for real-time API access:
POST /detect
curl -X POST "https://YOUR-STANDBY-URL/detect" \-H "Content-Type: application/json" \-d '{"text":"Your text to analyze here..."}'
GET /detect
$curl "https://YOUR-STANDBY-URL/detect?text=Your+text+to+analyze"
POST /detect/bulk
curl -X POST "https://YOUR-STANDBY-URL/detect/bulk" \-H "Content-Type: application/json" \-d '{"texts":["First text","Second text"]}'
Pricing
Pay-per-event pricing:
- $0.003 per text analyzed
- No charge for failed or insufficient text requests
- Bulk analysis charges per text in the batch
Limitations
- Statistical analysis only — Does not use ML models, so accuracy depends on text length and quality
- Best with 100+ words — Short texts (under 50 words) get low confidence ratings
- English-optimized — AI phrase detection is currently optimized for English text
- Not 100% accurate — No AI detector is perfect; use as one signal among many
- AI-assisted writing — Text that was AI-generated then heavily edited may score as "mixed"
Tips for Best Results
- Submit longer texts — 200+ words gives the most accurate results
- Submit raw text — Don't clean or format the text before analysis
- Multiple samples — Analyze several paragraphs from the same source for a pattern
- Use bulk mode — Compare multiple texts from the same author
Related Tools
- Email Validator API — Verify email addresses with MX, SMTP, and disposable checks
- Website Intelligence API — Detect technologies, analyze SEO, extract metadata
Browse all tools by george.the.developer on Apify.
