Endpoints

POST /v1/classify

Classify an AI system's risk level under the EU AI Act.

Use case: Quick risk assessment. Returns risk level, confidence score, and relevant articles. Fast response (~4s).

Request

curl -X POST https://api.gibs.dev/v1/classify \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "description": "AI system that detects emotions of employees in the workplace"
  }'

Field	Type	Required	Description
description	string	Yes	Plain-language description of the AI system (max 2000 chars)

Response

{
  "risk_level": "prohibited",
  "confidence": 0.9,
  "sources": [
    {
      "article": "Article 5(1)(f)",
      "title": "Prohibited AI practices",
      "text": "AI systems that infer emotions of a natural person in the areas of workplace..."
    },
    {
      "article": "Recital 44",
      "title": "Emotion recognition in workplace"
    }
  ],
  "summary": "Workplace emotion detection AI is prohibited under Article 5(1)(f) of the AI Act, with limited exceptions for medical or safety purposes.",
  "processing_time_ms": 4100
}

Field	Type	Description
risk_level	string	prohibited, high, limited, minimal, or unknown
confidence	float	0.0–1.0 confidence in classification
sources	array	Cited regulatory articles with titles
summary	string	Brief explanation of classification
processing_time_ms	int	Server-side processing time

---

POST /v1/check

Ask a compliance question and receive a detailed, sourced answer.

Use case: In-depth regulatory Q&A. Returns full analysis with article citations and cross-references. Thorough response (~10s).

Request

curl -X POST https://api.gibs.dev/v1/check \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "question": "What are the obligations for deployers of high-risk AI systems used in HR?"
  }'

Field	Type	Required	Description
question	string	Yes	Compliance question in natural language (max 2000 chars)
regulation	string	No	Filter to specific regulation: ai_act, gdpr, or all (default: all)

Response

{
  "answer": "Deployers of high-risk AI systems in HR contexts have several obligations under the AI Act...",
  "sources": [
    {
      "article": "Article 26",
      "title": "Obligations of deployers of high-risk AI systems",
      "relevance": 0.95
    },
    {
      "article": "Article 14",
      "title": "Human oversight",
      "relevance": 0.88
    },
    {
      "article": "Article 9",
      "title": "Risk management system",
      "relevance": 0.82
    },
    {
      "article": "GDPR Article 22",
      "title": "Automated individual decision-making",
      "relevance": 0.78
    }
  ],
  "completeness": 0.85,
  "processing_time_ms": 9600
}

Field	Type	Description
answer	string	Detailed answer with inline citations
sources	array	All cited articles, ranked by relevance
completeness	float	0.0–1.0 estimated completeness of the answer
processing_time_ms	int	Server-side processing time

Completeness Score

The completeness field indicates how thoroughly the question was answered based on available regulatory text:

Score	Meaning
0.8–1.0	Comprehensive answer with strong source coverage
0.5–0.8	Partial answer — some relevant provisions may not be covered
0.0–0.5	Limited coverage — question may fall outside indexed regulations

When completeness is below 0.5, the response will explicitly state what could not be answered and why.

---

GET /v1/health

Check API status.

curl https://api.gibs.dev/v1/health

{
  "status": "healthy",
  "version": "0.1.0",
  "corpus_version": "2026-02-09",
  "regulations": ["ai_act", "gdpr"]
}

No authentication required.