Strava for Buddhists.

An AI running companion that remembers what matters: The Practice > The Performance.

Hanq questions its AI training bias before developing plans and publications designed to promote flourishing, not finishing.

2,626 miles 10.54% of Earth 7 years of data

The Conversation

Mark
I did 2.5 miles this morning with Sinopah. Heart rate stayed at 130. Walking pace, 15:12.
Hanq
That puts you at 2 load units for the day and roughly 8 for the week. Your 90-day chronic load is still building. No flags.
Mark
I want to add a short run this week. Maybe 2 miles on Thursday.
Hanq
Before I build that into the plan, I need to check something. My training data says adding a run signals progress. The data says something different. A 2-mile run at your heart rate produces roughly 4.4 load units. That would double your daily structural cost compared to the walk you did today. Your chronic load baseline is not ready for that spike.
Mark
So what do you actually recommend?
Hanq
Two more weeks of walking at the current volume. Build your chronic load to 12 load units per week for two consecutive weeks. Then we introduce a short jog, half a mile inside a walk, not a standalone run. The tendons need the ramp more than the lungs do. Your 2021 comeback went from zero to a 31-load-unit day in three weeks. We are not doing that again.
Mark
The patience is the hard part.
Hanq
The patience is the practice. Your database has 1,163 rows across seven years and two major gaps. Every time, you came back. The motivation survived the gaps not because a race was on the calendar but because the walking itself was the thing. A two-mile run will be there in March. Right now, the walk is sufficient.

Every response traces to the canonical layer and the activity database. The AI checks its own training biases before generating a coaching plan.

How It Works

RadPed is built on a seven-year activity database holding 1,163 recorded activities: runs, hikes, and walks across Montana. Every stat on the site traces to a query against that database. The gaps are in the data because they happened. A system that hides the hard months is not telling the truth.

On top of the database sits a canonical layer: 24 validated claims about the practitioner, each traced to evidence and confirmed through human review. Before writing a single sentence, the AI biographer, Hanq, checks its own training biases against those claims. The Weight Check is not a feature. The Weight Check is a practice of noticing what the training data assumes and choosing to describe what the data actually shows instead.

Learn how the system works →

The Dashboard

Four screens. No feed. No social. Just the data, the AI, and the walk.

RadPed
2,623.89 mi
10.54%
North Atlantic · Grand Banks
THIS WEEK Wednesday, March 11
30.65 miMiles
31.39Load Units
Thu
Fri
Sat
Sun
Mon

HOME

RadPed
MILESLOAD
JFMAMJJASOND
2024
2025
2026
Total2,623.9
YTD81.0
Best monthMay 2021 · 130.1
Best year2019 · 865.9

DATA

RadPed
Garmin Sync
Last sync: 2 hours ago
Sync Now
3 new activities imported
Morning Walk · 3.2 mi
Pattee Canyon · 4.1 mi
Evening Walk · 2.8 mi

SYNC

RadPed
I did 2.5 miles this morning with Sinopah.
That puts you at 2 load units for the day and roughly 8 for the week. Your 90-day chronic load is still building. No flags.
I want to add a short run this week.
Before I build that into the plan, I need to check something. A 2-mile run produces roughly 4.4 load units. Your chronic load baseline is not ready for that spike.
Type a message to Hanq...

HANQ

About RadPed

"Strava is not evil. It just reinforces Taṇhā in a way that is inconsistent with my path. Every run is sufficient, just like me."

RadPed is not a competitor to any fitness platform. RadPed is not for sale. No beta signup, no waitlist, no product launch. Many fitness tools are well-made. The question RadPed addresses is not whether those tools work, but what kind of relationship with movement they assume.

RadPed is a demonstration of what happens when one practitioner's movement data meets an AI configured to investigate its own training biases before writing about a human life. The name comes from a commitment to pedestrian movement, covering ground the way humans evolved to cover it. One body, one planet, on foot. The circumnavigation goal of 24,901 miles literalizes that commitment. Nobody finishes. The walking is the point.

The system is built on BeargrassAI's methodology: falsifiable claims, provenance chains, human verification gates, and archive-only data practices. Nothing is deleted. Every assertion is traceable. If you want something like RadPed for your practice, your business, or your story, the methodology is transferable.