Why an independent cycling lab?
A note on making focused tools, showing the work and keeping the rider larger than the data.
Read noteÉchapper / The Lab
Échapper is an independent cycling lab. We build focused products that help riders understand more, decide better and stay in control—without turning every idea into another platform.
The working thesis
Échapper is not trying to become the place where every metric, conversation and ride must live. Each project should solve one understandable problem, work with the tools a rider already trusts and remain useful on its own.
Workbench
No fictional roadmap. This register shows projects only when there is real work to point to.
Working method
Start with a problem that exists outside the screen.
Find the smallest useful intervention and remove the rest.
Use it under real conditions and record what does not work.
Publish it when it earns a place—or stop before complexity becomes the product.
The Lab Standard
Not aspirations. These are the boundaries each Échapper product should make visible.
Sources stay visible.
Facts and interpretations remain separate.
Uncertainty is not hidden.
Important actions require deliberate approval.
Connected services remain replaceable.
A rider can disconnect and leave.
Products do not stay alive merely to fill a catalogue.
Independence
Échapper can work with services such as Intervals.icu, Claude and ChatGPT without pretending to be part of them. Those services remain separate, connections stay visible and a rider can remove access again.
Échapper is not affiliated with or endorsed by Intervals.icu, Anthropic, OpenAI or Garmin. Any future commercial relationship will be stated plainly.
See current integrationsFounder / Builder
Hi, I’m Andreas. I founded and build Échapper. I make the product decisions, test Training Companion with my own connected training data, write these pages and answer support.
Échapper is deliberately small, so questions and criticism reach the person doing the work. It does not claim to replace a qualified coach, medical professional or your own judgment.
From the Journal
A note on making focused tools, showing the work and keeping the rider larger than the data.
Read noteA complete guide to adding Échapper as a remote connector, checking the training context and reviewing controlled changes.
Read noteConnect ChatGPT to your training context and review explicitly approved changes.
Read noteThe rule