Syncflow tracks how long crumbs actually take vs. what the AI estimated. Over time, this builds a picture of your real pace and improves future plans.
Automatic Time Tracking
When you mark a crumb as done in Focus Flow, Syncflow automatically calculates the actualMinutes based on how long since you started working on that crumb. There is no timer to start or stop -- it is tracked passively in the background based on when the crumb became active and when you completed it.
Accuracy Percentage
Your dashboard shows an overall estimate accuracy percentage across all tracked crumbs. This compares the AI's estimated time to your actual completion time. An accuracy of 100% means the estimates perfectly match your real pace. Below 100% means tasks are taking longer than estimated; above 100% means you are finishing faster.
Tendency
Based on your history, Syncflow tells you whether you tend to underestimate (things take longer than expected) or overestimate (you finish faster than expected). This is useful for self-awareness and for calibrating how you think about time commitments in your work.
Template Learning
When you save a completed task as a template, the actual times you recorded replace the AI's original estimates. So the next time you use that template, every crumb's time estimate is based on your real data from the previous run. Over repeated uses, your templates become increasingly accurate and personalized.
How it works under the hood
Each crumb has an estimatedMinutes field (set by the AI or by you during refinement) and an actualMinutes field (calculated when you complete the crumb). The accuracy formula is:
accuracy = (estimatedMinutes / actualMinutes) * 100
Your overall accuracy is the average across all crumbs where both values exist.
Viewing estimates in the API
The GET /api/tasks/:id endpoint returns each crumb with both estimatedMinutes and actualMinutes. The GET /api/stats/digest endpoint includes aggregate time data. See the API Reference for details.