Spaces:
Sleeping
Sleeping
|
Download README_SPACES.md from webflow97/deployment-taskflow: direct link, hf CLI and curl.
- Browser
- Download file 1.27 kB
-
https://huggingface.co/spaces/webflow97/deployment-taskflow/resolve/main/README_SPACES.md
- Command line
-
hf download hf://spaces/webflow97/deployment-taskflow/README_SPACES.md
-
curl -L -o README_SPACES.md https://huggingface.co/spaces/webflow97/deployment-taskflow/resolve/main/README_SPACES.md
1.27 kB
metadata
title: Smart Escalation API
emoji: 🤖
colorFrom: blue
colorTo: green
sdk: docker
pinned: false
license: mit
Smart Escalation API
AI-powered L1 customer support system with intelligent escalation.
Features
- RAG-based retrieval using sentence-transformers and FAISS
- Intelligent escalation to human agents
- Google Gemini LLM integration
- FastAPI backend with automatic documentation
API Endpoints
GET /- API informationGET /health- Health checkPOST /ask- Submit customer questions
Environment Variables
Configure these in your Space settings:
GOOGLE_API_KEY- Your Google Gemini API key (required)ARTICLES_DIR- Path to help articles (default: data/articles)EMBEDDING_MODEL- Sentence transformer model (default: all-MiniLM-L6-v2)RELEVANCE_THRESHOLD- Minimum similarity score (default: 0.5)TOP_K_CHUNKS- Number of chunks to retrieve (default: 3)LLM_MODEL- Gemini model name (default: gemini-1.5-flash)CORS_ORIGINS- Allowed CORS origins
Usage
curl -X POST https://YOUR_USERNAME-escalation-api.hf.space/ask \
-H "Content-Type: application/json" \
-d '{"question": "How do I reset my password?"}'
Documentation
Visit /docs for interactive API documentation.