32 lines (21 loc) · 765 Bytes

AI-Server

Conversation-based risk decision support system with a deterministic pipeline.

Run

Install dependencies: pip install -r requirements.txt
Start API: uvicorn app.main:app --reload

API

Endpoint: POST /api/analyze
Base URL: http://ai-server:8000
Request: JSON
Response: JSON

Request Rules

platform: INSTAGRAM 또는 TELEGRAM만 허용
messages[].type: TEXT 또는 URL
URL 메시지는 OCR로 텍스트를 추출해 분석 파이프라인에 합쳐 처리

Swagger

http://localhost:8000/docs

환경변수

OPENAI_API_KEY + OpenAI API 키 필요
OPENAI_OCR_MODEL (기본값: gpt-4o-mini)
OCR_DOWNLOAD_TIMEOUT_SECONDS (기본값: 10)
OCR_MAX_IMAGE_BYTES (기본값: 5000000)