Security in MakerAI

• System Architecture

MakerAI operates based on a modern architecture:

• Next.js (web application)
• Node.js server managed by PM2
• Reverse proxy (nginx, HTTPS)
• PostgreSQL database (Docker)
• Hosting on VPS servers

• Data Protection

• all connections are encrypted (HTTPS)
• the database is not publicly accessible
• system access is restricted and controlled

• AI Processing

• data is processed solely for problem analysis
• we minimize the scope of shared data
• we do not use data to train models without consent

• Matching and storing results

Recommendations combine data in PostgreSQL with public sources:

• registered expert profiles (including profile embeddings — defined in the Privacy Policy under Scope and Sources of Data) and metrics are stored in the database
• GitHub matching results (candidate lists, local/global splits) are cached on the request with a freshness time window
• GitHub is queried when building the candidate pool and when topping up results live, with timeouts to avoid long waits

• Logs and Monitoring

To ensure system stability:

• we monitor application errors
• we analyze server logs
• we detect abuse and anomalies

• Access and Authorization

• access to administrative panels is restricted
• role-based access control is applied (user / admin / expert)
• user sessions are secured

• Protection Against Abuse

• the system detects suspicious activity
• rate limiting is applied
• APIs are protected against unauthorized access

• Backups

• we perform regular data backups
• the system is prepared for data recovery

• Vulnerability Reporting

If you have found a security vulnerability: podkowa71@gmail.com

Thank you for responsible disclosures.

• Our Approach

We build MakerAI as an:

AI-powered problem solving platform,

with a strong focus on security, privacy, and reliability.

Last updated: 2026-05-03