The Deal

Reviewed by TypeFuel Lab. Last updated July 18, 2026. Plain-language Research Preview privacy boundary.

You give

Timing patterns and personal baselines. We never see the content of what you type - only the rhythm.

You get

Free desktop preview access, a personal baseline view, and clear updates on what the signal gets right or wrong.

The honest summary

During the Research Preview, aggregate data leaves your computer so we can test whether the signal works. If validation earns a v1.0 score, the goal is local-only detection.

What data does TypeFuel collect?

During the Research Preview, TypeFuel collects per-minute aggregates that help test whether a cognitive load tracker can work from work rhythm alone. The desktop app summarizes keyboard timing, correction patterns, research-only cursor rhythm metrics, active and idle rhythm, plus cognitive load summaries from the personal baseline. It does not store raw keystrokes, typed content, app names, window titles, browser URLs, screen contents, file names, contacts, microphone, or camera. Preview data exists to test the signal. Founder reservation email inside the app stays separate from behavioral research data. In v1.0, if validation earns it, the goal is local detection so behavioral data stays on the user's computer. Until then, preview participants can pause collection, export what they have contributed, withdraw, and request deletion.

Keyboard timing and correction patterns. Research-only cursor rhythm metrics. Personal baseline and cognitive load summaries.

What we send is computed once per minute, not streamed live.

1. Work content stays
2. One-minute timing summary
3. Research validation

Exactly what we don't collect

Content of typing. Apps. Websites. Window titles. Browser URLs. Screen contents. What you clicked on. Microphone. Camera. Names of people you communicate with.

People comparing privacy-friendly work tools often start with a practical constraint: understand their work without screenshots, app names, URLs, or typed content. TypeFuel approaches that problem with timing summaries and personal baselines. It is not a time tracker and does not reconstruct how time was spent. Read how TypeFuel measures cognitive load from those private timing signals.

Collected

• ✓ Keyboard timing and correction patterns
• ✓ Cursor rhythm metrics for research
• ✓ Personal baseline and cognitive load summaries

Never collected

• ✗ Content of what you type
• ✗ Apps and websites you use
• ✗ What you clicked on
• ✗ Screen contents
• ✗ Microphone
• ✗ Camera
• ✗ Names of people you communicate with

Where it goes

Aggregated timing patterns and baseline summaries go to our research server. Access is restricted to two named researchers. Retention is 12 months from collection. You can delete everything at any time and we wipe it within 30 days.

A stable per-user ID lets us compare you against your own past. We don't tie it to anything personally identifiable.

Your controls

• Pause collection any time from the menu bar / system tray.
• Export everything you've contributed in one click.
• Delete everything you've contributed in one click.
• Withdraw and we wipe your data within 30 days.

What we'll do with it

Test the detection signal. Publish aggregate findings. Nothing else: no selling, no advertisers, no third-party sharing, no individual reports. See the Roadmap for what TypeFuel will and will not become after the preview, or return to the measurement and validation methodology.