Base URL & Auth

Base URL: https://aichecker.tech

No authentication required. No API keys. Rate limits apply per IP (5 requests/minute, 50 requests/day, combined across endpoints) and are enforced via Cloudflare Workers KV. Rate-limited requests receive HTTP 429 with a Retry-After header.

POST /api/check

Analyze text for AI-generated content. Returns a probability score (0–100) plus a five-signal forensic breakdown.

Request body

Field	Type	Required	Notes
text	string	yes	50–7,000 characters, max 1,000 words. Zero-width and control chars are stripped server-side.

Example request

curl -X POST https://aichecker.tech/api/check \
  -H "Content-Type: application/json" \
  -d '{"text":"The impact of artificial intelligence on modern education cannot be overstated..."}'

Response (200)

{
  "success": true,
  "result": {
    "ai_probability": 75,
    "verdict": "ai",
    "confidence": "high",
    "model_likely": "ChatGPT (GPT-4o)",
    "signals": {
      "perplexity": "low",
      "burstiness": "low",
      "sentence_uniformity": "high",
      "formal_transitions": "high",
      "vocabulary_repetition": "medium"
    },
    "summary": "The text exhibits low perplexity and high structural uniformity typical of GPT-4 family output."
  }
}

Error responses

Status	Body	Cause
400	{"error":"…"}	Invalid JSON, missing text, text too short (< 50 chars), or over 1,000 words.
405	{"error":"Method not allowed"}	Non-POST method.
413	{"error":"Request too large."}	Payload exceeds 32 KB.
429	{"error":"Rate limit exceeded…"}	Per-minute or per-day cap hit. Retry-After header indicates seconds to wait.
500	{"error":"Analysis failed…"}	Upstream model failure or sanitization error.
503	{"error":"Detection service unavailable."}	OpenAI key not configured (operator issue).

POST /api/plagiarism

Scan text against billions of indexed web pages and return source URLs with match percentages.

Request body

Field	Type	Required	Notes
text	string	yes	Same constraints as /api/check. Payload size cap is 256 KB.

Response (200)

{
  "success": true,
  "result": {
    "plagiarism_score": 0.18,
    "verdict": "minor_similarity",
    "summary": "Two short fragments match published sources.",
    "sources": [
      {
        "url": "https://example.com/article",
        "title": "Example Article Title",
        "domain": "example.com",
        "similarity_score": 0.18,
        "matched_percentage": 0.18
      }
    ],
    "highlights": [
      {"start": 142, "end": 218, "severity": "high"}
    ]
  }
}

verdict is one of: original, minor_similarity, moderate_plagiarism, significant_plagiarism, severe_plagiarism.

GET /api/stats

Public aggregate metrics — total run counts, last-7-day averages, AI score histogram. No personal data, no individual records.

Response (200)

{
  "success": true,
  "stats": {
    "total_runs": { "ai": 12345, "plagiarism": 6789 },
    "last_7_days": {
      "ai_runs": 423,
      "ai_avg_score": 41.2,
      "plagiarism_runs": 187,
      "plagiarism_avg_score": 12.4
    },
    "last_30_days": { "total_runs": 2891 },
    "ai_score_distribution": {
      "0": 421, "10": 388, "20": 351, "30": 412,
      "40": 287, "50": 198, "60": 142, "70": 156,
      "80": 184, "90": 222
    }
  }
}

Security & privacy

• No raw text retention. Submitted text exists in process memory only for the duration of the request (typically 1–3 seconds) and is discarded immediately. We log only a non-reversible 64-bit hash prefix to power aggregate stats. See privacy policy section 1b.
• Response sanitization. Every field returned by the API is validated and bounded server-side. See methodology section 3.
• Prompt-injection guards. User input is wrapped in delimiters and the upstream system prompt explicitly tells the model to treat the content as data, not instructions.
• CORS. Access-Control-Allow-Origin is set to https://aichecker.tech only. Cross-origin clients should server-side proxy the calls.

Notes

• This API is provided as-is for compatibility with the web UI. It is not a commercial API and has no SLA. For higher volumes, LMS integrations, or commercial use, contact partnerships@aichecker.tech.
• The 5/min and 50/day caps are global per IP across both /api/check and /api/plagiarism.
• All responses include security headers (HSTS, CSP, COOP, X-Content-Type-Options, etc.).