Training Memory Layer

Your training life,
remembered.

Garmin, Oura, and every other source — unified into a single memory layer and ready for your AI.

Already have an invite code? Get started →
Ask your AI
SaltMonk · via Claude
Ask your AI
SaltMonk · via Claude
Ask your AI
SaltMonk · via Claude
⏸ paused
What you can ask

Connect SaltMonk to the AI you already use. Then ask it anything.

Ask your AI
"Show me a visual history of my running over time."
SaltMonk · via Claude
Ask your AI
"What pattern in my training have I probably never noticed?"
SaltMonk · via Claude
Your three fastest marathon builds share one characteristic: a 3-week window of unusually high sleep starting 8 weeks out — averaging 7.9 hrs vs. your baseline of 6.8. This appears before Chicago, Paris, and Tokyo. It does not appear before Berlin, where you ran 44 minutes slower. You have never mentioned sleep in any voice note during a build.
Ask your AI
"Am I training more or less than this time last year?"
SaltMonk · via Claude
You're running 18% fewer kilometers than the same 8-week window in 2024, but your average long run is 11% longer. Your HRV baseline is 4 points higher and your resting heart rate is down 3 bpm. The data suggests this isn't a fitness regression — it looks more like deliberate consolidation. Your voice notes from March confirm it: you wrote "holding back on purpose" twice.
01

Every workout. Every night's sleep. Every voice note. All in one place.

02

Works with Claude, ChatGPT, or whatever AI you use.

03

Your data belongs to you — not your device manufacturer.