How harishpullagurla/codecred_ai works
CodeCred AI is a Flask web application for comparing skills claimed in a resume with evidence from a candidate’s GitHub repositories. A user submits a GitHub username and PDF resume; the Flask route extracts resume text, fetches and scans repositories for languages and technologies, then asks Gemini to generate an HTML verification report. The report is returned to the frontend. The sampled backend shows these calls; the frontend source was not sampled, so its submission and result-display connections are unverified.
Main entry points: app.py Flask routes · index.html Web interface · pdf_service.py Resume extraction · github_service.py GitHub repository scan
Components 9
Web application
- Web interface
templates/index.html - Flask routes
app.py
Evidence collection
- Resume extraction
services/pdf_service.py - GitHub repository scan
services/github_service.py
Verification analysis
- Skill verification report
services/ai_service.py
Other components
- Candidate or reviewer
- GitHub REST API
- PDF reader library
- Google Gemini API
Connections 9
- Flask routes → Web interface: renders page
app.py - Candidate or reviewer → Web interface: uses
- Web interface → Flask routes: submits analysis
- Flask routes → Resume extraction: extracts resume
app.py - Resume extraction → PDF reader library: reads PDF
services/pdf_service.py - Flask routes → GitHub repository scan: fetches evidence
app.py - GitHub repository scan → GitHub REST API: requests repository data
services/github_service.py - Flask routes → Skill verification report: requests report
app.py - Skill verification report → Google Gemini API: generates report
services/ai_service.py