Social Listening AI Analysis API

Python FastAPI backend that provides LLM-powered insights and recommendations for the Social Listening dashboard. Uses Google Gemini to analyse Kalventis vs GSK competitive data.

Endpoints

Method	Path	Description
GET	/api/health	Health check
POST	/api/v1/analysis	Generate AI analysis from dashboard snapshot

Prerequisites

• Python 3.12+
• A Google Gemini API key

Setup

1. Clone / navigate to this directory

cd D:\fastapi_all\python-social-listening

2. Create and activate a virtual environment

python -m venv .venv

# Windows
.venv\Scripts\activate

# Mac / Linux
source .venv/bin/activate

3. Install dependencies

pip install -r requirements.txt

4. Configure environment variables

Copy .env.example to .env and fill in your Gemini API key:

cp .env.example .env

.env contents:

GEMINI_API_KEY=your_gemini_api_key_here
GEMINI_MODEL=gemini-2.5-flash

Running

Development (with auto-reload)

uvicorn main:app --port 8000 --reload

Production

uvicorn main:app --host 0.0.0.0 --port 8000 --workers 2

The API will be available at http://localhost:8000.

Check it's running:

curl http://localhost:8000/api/health
# {"status":"ok"}

Interactive API docs: http://localhost:8000/docs

Running with Docker

Build the image

docker build -t social-listening-api .

Run the container

docker run -p 8000:8000 --env-file .env social-listening-api

Connecting to the Frontend

The Next.js frontend (at social-listening-monitoring/) calls this backend via the /api/analysis proxy route. Make sure this backend is running on port 8000 before clicking Generate Analysis on the dashboard overview page.

If you need to run it on a different port, set ANALYSIS_API_URL in the Next.js .env:

ANALYSIS_API_URL=http://localhost:9000