Legal

Data Handling & Security

Last Updated: March 2025

At Ninchi, we believe developer tools should respect the sensitivity of source code. Ninchi is designed to be code-aware, not code-retentive. We process code only when necessary and avoid storing it whenever possible.

How Ninchi Works

When a pull request is opened:

  1. Ninchi temporarily accesses the relevant code changes
  2. A targeted question is generated based on the diff
  3. The developer provides an answer
  4. The system evaluates the response
  5. Raw code is discarded after processing

What We Do NOT Store

Full repository contentsLong-term pull request diffsProprietary source code

What We DO Store

We store only derived information:

Questions & responsesEvaluation scoresMetadata (language, category, difficulty)Usage analytics

This lets us provide meaningful insights without retaining sensitive code.

Security Measures

Infrastructure

  • TLS encryption for all data in transit
  • Encryption at rest for stored data
  • Private networking for backend services
  • Secure cloud infrastructure (AWS)

Access Controls

  • Role-based access controls
  • Secure secret management
  • Least-privilege service accounts

AI & Model Usage

  • We do not train models on customer code
  • We do not share code across customers
  • We focus on patterns and understanding — not proprietary logic

Aggregated Insights

Ninchi may surface insights such as common areas where developers struggle, trends across programming languages, and difficulty patterns in code comprehension. These insights are:

AggregatedAnonymizedNot tied to specific codebases

Designed for Trust

Ninchi is built with the assumption that your code is one of your most valuable assets. Our architecture reflects that. We store the minimum required to operate the Service — nothing more.

Questions

Have a security question or concern? Contact us at support@ninchi.ai.