How miteshpanda/truelens-ai works
TrueLens AI is a multimodal misinformation-analysis application. A user selects text or image/video analysis in the web UI; the corresponding page sends a request through the API client to FastAPI, which returns a verdict, confidence, and supporting details for display. Text analysis combines linguistic scoring with optional web verification; media analysis uses image/video detection, with sampled video-frame results available to the UI. The README also describes multi-modal fusion, but the sampled implementation does not establish a fusion component, so that connection is not asserted. Some backend route-to-model wiring is omitted from excerpts.
Main entry points: App.jsx App navigation · TextAnalysis.jsx Text analysis page · MediaAnalysis.jsx Media analysis page · main.py Detection API
Components 17
Web application
- App navigation
frontend/src/App.jsx - Home page
frontend/src/pages/Home.jsx - Navigation bar
frontend/src/components/Navbar.jsx - Text analysis page
frontend/src/pages/TextAnalysis.jsx - Media analysis page
frontend/src/pages/MediaAnalysis.jsx - Verdict display
frontend/src/components/ResultCard.jsx - Confidence meter
frontend/src/components/ConfidenceMeter.jsx - Web evidence panel
frontend/src/components/WebEvidencePanel.jsx - Frame timeline
frontend/src/components/FrameTimeline.jsx
API boundary
- API client
frontend/src/api.js - Detection API
backend/main.py
Text analysis
- Fake-news analysis
backend/models/fake_news_model.py - Text preprocessing
backend/utils/preprocess.py - Claim verification
backend/utils/web_verifier.py - Wikipedia API
Media analysis
- Deepfake detection
backend/models/deepfake_model.py
Other components
- User
Connections 20
- User → App navigation: uses
README.md - App navigation → Navigation bar: renders
frontend/src/App.jsx - App navigation → Home page: renders
frontend/src/App.jsx - App navigation → Text analysis page: renders
frontend/src/App.jsx - App navigation → Media analysis page: renders
frontend/src/App.jsx - User → Text analysis page: submits text
frontend/src/pages/TextAnalysis.jsx - User → Media analysis page: uploads media
frontend/src/pages/MediaAnalysis.jsx - Text analysis page → API client: requests text
frontend/src/pages/TextAnalysis.jsx - Media analysis page → API client: requests media
frontend/src/pages/MediaAnalysis.jsx - API client → Detection API: posts predictions
frontend/src/api.js - Detection API → Fake-news analysis: calls text model
backend/main.py - Fake-news analysis → Text preprocessing: uses token extraction
backend/models/fake_news_model.py - Fake-news analysis → Claim verification: verifies claims
backend/models/fake_news_model.py - Claim verification → Wikipedia API: queries
backend/utils/web_verifier.py - Detection API → Deepfake detection: invokes media analysis
backend/main.py - Text analysis page → Verdict display: renders result
frontend/src/pages/TextAnalysis.jsx - Media analysis page → Verdict display: renders result
frontend/src/pages/MediaAnalysis.jsx - Verdict display → Confidence meter: renders meter
frontend/src/components/ResultCard.jsx - Text analysis page → Web evidence panel: renders evidence
frontend/src/pages/TextAnalysis.jsx - Media analysis page → Frame timeline: renders frames
frontend/src/pages/MediaAnalysis.jsx