Prodini Code Navigator Privacy Policy
Last updated: March 2026
Privacy Summary
Prodini Code Navigator connects ChatGPT to your organization's Git repositories to enable AI-powered code search and repository navigation.
Data We Process
The service processes only the data necessary to provide repository search and exploration features, including:
- Repository metadata (repository names, file paths, branches)
- Source code from administrator-selected repositories
- Commit messages and pull/merge request metadata
- User account information (email and organization)
AI Processing
Prodini uses OpenAI's Embedding API (text-embedding-3-small) to generate semantic vectors that enable repository search.
Only truncated repository text required for embedding generation is sent to OpenAI.
This data is not used to train OpenAI models.
Security
- Git credentials are encrypted using AES-256-GCM.
- All traffic is encrypted via TLS.
- All repository data is isolated per organization (tenant).
Data Control
Organizations control which repositories are indexed and can remove integrations or delete indexed data at any time.
For full details, see the complete Privacy Policy below.
1. Introduction
Prodini Code Navigator is an AI development assistant that connects ChatGPT to your organization's Git repositories to enable semantic search and repository exploration.
This page describes the privacy practices specific to the Prodini Code Navigator MCP connector, which allows AI assistants such as ChatGPT to access repository information through secure read-only tools.
The service operates in a read-only mode with respect to connected repositories and does not modify source code, commits, or pull requests.
This document supplements the main Prodini Privacy Policy.
2. Information We Process
Prodini Code Navigator processes only the information required to enable repository indexing and search capabilities.
Account Information
When users authenticate through the Prodini platform we may process:
- Email address
- Organization identifier
- User role within the organization
Passwords are stored only as secure cryptographic hashes.
3. Repository Data
When an administrator connects a Git provider, the service may retrieve data from selected repositories including:
- Repository metadata
- Repository file structure
- Source code files
- Commit messages
- Pull or merge request titles and descriptions
- Commit author metadata contained within the repository history
Only repositories explicitly selected by the organization administrator are indexed.
4. AI Processing
Prodini Code Navigator uses OpenAI's Embedding API to generate semantic embeddings used for repository search.
Embedding model used: text-embedding-3-small
Data sent for embedding may include truncated portions of:
- Source code
- Commit messages
- Pull request descriptions
Important limitations:
- Data is used only for embedding generation
- The service does not send customer data to LLM completion models
- Data sent through the OpenAI API is not used for model training
Embeddings generated from repository data are stored in the platform database and used only to enable semantic search functionality.
Prodini does not generate AI responses itself. ChatGPT generates responses using the tool results returned by the service.
5. Third-Party Services
The service relies on a limited set of infrastructure providers.
| Provider | Purpose |
|---|---|
| OpenAI | Embedding generation for semantic search |
| Google Cloud Platform | Infrastructure hosting |
| GitHub | Repository access using customer credentials |
| GitLab | Repository access using customer credentials |
These providers process data only as required to provide their services.
6. Security
Prodini implements several security measures including:
- TLS encryption for all network communication
- AES-256-GCM encryption for stored integration credentials
- Tenant-level data isolation
- Role-based access control
- Access logging and operational monitoring
Git credentials are encrypted and decrypted only when required to access repository APIs.
7. Data Retention
Repository data and embeddings are stored only while the repository integration remains active.
Customers may:
- Remove repositories from indexing
- Disconnect integrations
- Delete indexed data
When an integration is removed, associated repository data and embeddings are deleted.
Backup copies may temporarily retain encrypted data according to infrastructure retention policies.
8. User Rights and Data Control
Organizations control the repositories connected to the service.
Administrators may:
- Connect or disconnect repositories
- Delete repository indexes
- Rotate integration credentials
- Request data export
- Request account deletion
Upon account termination, customer data is removed from the production system.
9. Logging
The service stores limited operational logs required for reliability and security.
Logged data may include:
- Tool execution metadata
- Request timestamps
- System performance information
The service does not store:
- ChatGPT prompts
- AI responses generated by ChatGPT
10. Contact
If you have questions about the privacy practices of Prodini Code Navigator, please contact: support@prodini.ai