Legal

Data Usage & Privacy Summary

Updated: May 22, 2026

If you have questions about how Codira handles your information, contact us at support@codira.ai. Additional details regarding data collection and processing can be found in our complete Privacy Policy.

Below is an overview of how data is managed depending on the privacy settings selected within Codira.

Privacy Mode Enabled

When “Privacy Mode” is turned on in Codira:

  • Zero-retention policies are applied with supported ai model providers.
  • Your source code and project content are not used for ai training by Codira or third parties.
  • Certain encrypted or temporary processing data may still be stored briefly to support platform functionality and advanced editor features.

Your code remains private and is excluded from model training workflows.

Privacy Mode Disabled

If you disable Privacy Mode:

  • Codira may collect and retain prompts, editor interactions, source code snippets, project metadata, and related telemetry to improve product performance and enhance ai capabilities.
  • This information may also be used for internal model improvement and training purposes.
  • When you intentionally select third-party ai models, limited prompts and telemetry may be shared with those providers as necessary to fulfill your request.

Certain infrastructure and inference partners may temporarily process or cache inputs and outputs for operational optimization and performance tuning. Any temporary retention is limited and removed after processing is complete.

Additional details regarding provider-specific handling practices are available on our Security & Trust documentation pages.

Legacy Privacy Mode

Users operating under “Privacy Mode (Legacy)” receive:

  • Zero data retention protections;
  • No permanent storage of code content;
  • No ai training using your code or project data by Codira or external providers.

Additional Information

API Key Usage

Even when you connect your own API key, requests continue to pass through Codira’s infrastructure in order to assemble prompts, coordinate context handling, and provide editor functionality.

Codebase Indexing

If you enable repository indexing:

  • Your codebase may be securely uploaded in segmented chunks for embedding generation and semantic indexing.
  • Plaintext source content used during embedding generation is discarded immediately after processing.
  • Embeddings, hashes, filenames, and related metadata may remain stored to support search and navigation features.

Temporary File Caching

To improve responsiveness and reduce bandwidth usage:

  • Codira may temporarily cache encrypted file contents during active requests.
  • Encryption keys are generated client-side and exist only for the duration of request processing.
  • Cached data is ephemeral, automatically removed, and never used for ai training while Privacy Mode is enabled.

Historical Account Note

For accounts created before October 15, 2025, prompts and telemetry are not shared with third-party model providers under the previous data-sharing configuration.

Data handling questions? Email support@codira.ai or see the Privacy Policy for full details.