SearchEval

Influencer Trust SaaS
A FastAPI-based SaaS platform for analyzing influencer content legitimacy and detecting fraud indicators like undisclosed sponsorships.

Built Features
🎥 Video Transcription Service
Endpoint: POST /api/v1/transcribe
Transcribes video content using OpenAI Whisper/LLM
Extracts video metadata via YouTube API integration
Validates channel profiles and partnership history
Generates legitimacy scores based on transcription and profile data
Stores transcriptions with vector embeddings in Qdrant
📊 Content Analysis Service
Endpoint: POST /api/v1/analyze
Analyzes content from URLs for legitimacy
Scores content validity (0-1 scale) using LLM
Generates AI-powered summaries of analyzed content
Detects fraud indicators (e.g., undisclosed sponsorships)
Retrieves related links and articles
🔍 Vector Search & RAG (Retrieval-Augmented Generation)
Qdrant Integration: Vector database for semantic search
Similar Case Search: Finds similar cases using cosine similarity
Two Collections:
analysis_results: Stores content analysis with embeddings
transcription_results: Stores video transcriptions with embeddings
Embedding Generation: Uses OpenAI text-embedding-ada-002 (1536 dimensions)
✅ Profile Validation
Channel profile verification
Partnership history checking
Fraud record detection
Related links discovery
🤖 LLM Services
Content scoring and summarization
Transcription processing
Embedding generation
Legitimacy assessment
🗄️ Data Storage
Qdrant Vector Database: For semantic search and similarity matching
Vector Embeddings: 1536-dimensional embeddings for all content
Metadata Storage: Stores scores, summaries, links, and validation results
🔌 API Integrations
YouTube Data API: Video metadata extraction
OpenAI API: LLM, Whisper transcription, and embeddings
ElevenLabs API: (Configured for future use)
🏗️ Infrastructure
FastAPI: Modern, fast web framework
Docker Compose: Qdrant service containerization
SQLAlchemy & Alembic: Database migrations support
Pydantic: Data validation and serialization
RESTful API: Well-structured API endpoints with OpenAPI documentation
🧪 Testing
Test suites for analysis service
Test suites for transcription service
Pytest framework integration
Tech Stack
Framework: FastAPI 0.104.1
Vector DB: Qdrant
LLM/Embeddings: OpenAI (GPT-3.5-turbo, Whisper, text-embedding-ada-002)
Database: SQLAlchemy 2.0.23 with Alembic migrations
Validation: Pydantic 2.5.0
HTTP Client: httpx 0.25.2
Testing: pytest 7.4.3
Getting Started
Install dependencies:
pip install -r requirements.txt
Set up environment variables in .env:

qdrant_url
qdrant_api_key
elevenlabs_api_key
youtube_api_key
openai_api_key
Start Qdrant with Docker Compose:

docker-compose up -d
Run the application:
python run.py
The API will be available at http://localhost:8000

API Documentation
Once the server is running, visit:

Swagger UI: http://localhost:8000/docs
ReDoc: http://localhost:8000/redoc
Notes
TODO: Replace OpenAI with Mistral for analysis service
Current implementation uses OpenAI for LLM, transcription, and embeddings