README.md

metadata

title: LangChain Chat API
emoji: 💬
colorFrom: blue
colorTo: purple
sdk: docker
pinned: false
app_port: 7860

LangChain Chat API

A production-ready Chat API powered by LangChain and HuggingFace Inference Endpoints. This lightweight FastAPI application proxies user queries to advanced Open-Source Large Language Models (like DeepSeek-R1-0528 or Qwen3) using the HuggingFaceEndpoint class, avoiding the need for expensive local GPU inference.

Features

• LangChain Integration: Uses PromptTemplate and ChatHuggingFace with structured outputs to cleanly format and return model responses.
• Lightweight Architecture: Eliminates the need for massive PyTorch or Transformers dependencies by offloading inference to Hugging Face.
• FastAPI Backend: Provides robust, asynchronous request handling with CORS middleware and global exception management.

---

Setup and Execution

1. Requirements

Since the application offloads inference to the Hugging Face API, you must provide your Hugging Face API Token.

1. Get a Hugging Face API Token from your Hugging Face Settings.
2. Set it as an environment variable (or as a Secret if deploying to Hugging Face Spaces):

export HUGGINGFACEHUB_API_TOKEN="hf_your_token_here"

2. Running with Docker

The provided Dockerfile builds a highly-optimized, slim Python image.

docker build -t chat-api .
docker run -p 7860:7860 -e HUGGINGFACEHUB_API_TOKEN="your_token" chat-api

3. Running Locally

# Clone the repository and navigate to the project directory
cd multimodal-rag

# Install Python dependencies
pip install -r requirements.txt

# Run the FastAPI application
python app/main.py

---

Example API Usage

1. Root / UI Check

When accessing the Space UI directly in a browser, the application returns a status payload pointing to the endpoints.

curl -X GET "http://localhost:7860/"

2. Health Check

curl -X GET "http://localhost:7860/api/health" 

3. Chat Endpoint

Submit a question to the LLM. It returns a structured JSON answer containing the answer and justification.

curl -X POST "http://localhost:7860/api/chat" \
     -H "Content-Type: application/json" \
     -d '{
           "question": "What is LangChain?"
         }'

Response Example:

{
  "answer": "LangChain is an open-source framework designed to simplify the creation of applications using large language models.",
  "justification": "The user asked for a definition of LangChain, which this provides concisely."
}