Cross-platform AI analysis that finds patterns in your data and delivers actionable recommendations every week.
Every week, xeve feeds your productivity, coding, health, music, and GitHub data into Claude and generates a personalized insight report. It discovers patterns like "you code 40% more after 7+ hours of sleep" and suggests concrete changes to optimize your workflow.
The AI analyzes data from every connected source — apps, coding, GitHub, Spotify, health, and location — to find correlations a human would miss.
Identifies recurring patterns like your most productive time of day, the effect of sleep on output, or how music listening correlates with focus.
Each insight includes specific, actionable suggestions. Not just "you should sleep more" but "shifting bedtime 30 minutes earlier correlates with 25% more coding output."
Insights are generated every Sunday night, giving you a fresh analysis at the start of each week. You can also trigger on-demand generation.
Each weekly report compares your current metrics against previous weeks, highlighting improvements, regressions, and emerging trends.
AI insights are generated by a Supabase Edge Function that queries the daily_summaries table for the past 7 days, along with raw data from coding_sessions, github_activity, spotify_history, and health_samples. The data is formatted into a structured prompt and sent to Claude via OpenRouter. The response is parsed and stored in the ai_insights table with user_id, content (markdown), and generated_at. The function runs on a weekly pg_cron schedule and can also be triggered manually from the dashboard.
Insights are generated by Claude (Anthropic) via OpenRouter. The model analyzes structured data from your dashboard — it never sees raw window titles or personal content, only aggregated metrics and categories.
The AI can generate insights with as little as one day of data, but the analysis improves significantly with a full week. After 2-3 weeks of data, correlations and trend detection become highly accurate.
Yes. AI insights are opt-in — they will not generate unless you have data flowing and the feature enabled. You can disable them at any time from your dashboard settings.
early access
free during early access. no credit card required. install the mac app and your dashboard is live in seconds.
19 auto-computed metric pairs with Pearson analysis and plain-English interpretations.
learn more →001Zero-config macOS menu bar app that tracks every application you use, window title, and category — passively, in the background.
learn more →005BLE heart rate from Whoop, Polar, and Garmin. HealthKit steps, sleep, active energy, and distance from iOS.
learn more →