Data Use & Privacy Overview

Last updated: Oct 20, 2025

If you have any questions or feedback, please email us at support@cruxtro.com.

For more information about how we collect, use, and protect your personal data, please see our full Privacy Policy.

---

How Cruxtro Uses Your Data

Cruxtro is designed to help product managers and teams connect insights, reduce context-switching, and act faster — all while keeping your data secure and under your control.

Here’s how your data is handled based on your workspace’s privacy settings and AI preferences:

---

If you enable “Private Mode”

• Zero data retention is enabled for our AI and model providers.
• Your workspace data, feedback, and product documents are not used for model training by Cruxtro or any third party.
• Cruxtro may temporarily store metadata (e.g., timestamps, document references, or anonymized insight IDs) to enable collaboration features, but this data is not accessible to model providers.
• Workspace content, feedback, and insights remain encrypted and are processed only to deliver your requested outputs.

In short: No model training, no long-term storage, and full data privacy.

---

If you disable “Private Mode”

When “Private Mode” is off, you allow Cruxtro to use limited workspace and feedback data to improve our AI features and product intelligence.

We may process:

• Product-related inputs (e.g., feedback summaries, insight tags, metrics)
• AI prompts and responses (to improve quality and accuracy)
• Interaction telemetry (e.g., feature usage, click and search patterns)
• Workspace context (e.g., project metadata or linked data sources)

This data helps us enhance features like context linking, flow intelligence, and automated product insights.

We may share limited, anonymized data with our trusted model providers (such as OpenAI or Anthropic) only when you explicitly select their models.

Model provider access:

Some inference partners may temporarily store prompts and responses to improve inference performance.

This data is automatically deleted after use and never used to train third-party models.

---

If you use “Legacy Private Mode”

• Zero data retention applies (as above).
• None of your workspace, product, or feedback data will be used for AI training or improvement.

---

Other Important Notes

• All API and model calls — even if you use your own model key — route through Cruxtro’s secure backend to ensure consistent prompt handling and privacy control.
• If you connect external sources (e.g., Notion, Jira, or Slack), Cruxtro only accesses data you explicitly grant permission for, and all syncs can be revoked anytime.
• When you enable workspace indexing, Cruxtro may temporarily upload small encrypted chunks of your data to compute semantic embeddings.
  • These plaintext chunks are deleted immediately after processing.
  • The resulting embeddings (mathematical vectors, not readable content) and metadata (like file names or hashes) may be securely stored to enable search, summarization, and context mapping.
• We temporarily cache certain content to improve performance and reduce latency.
  • Cached data is encrypted using unique, session-specific keys and automatically deleted after the session ends.
  • Cached content is never permanently stored, shared, or used for AI training when Private Mode is active.

---

Summary

---

Your trust is core to Cruxtro’s mission.

We’re building AI-native tools that reduce cognitive overload — not increase data risk.

To learn more about our infrastructure and security practices, visit our Security Overview.

---