Greeting Vital MTBer,
Over the years a lot of riders have asked me for help with suspension setup, and I kept seeing the same problem: they know the bike feels wrong, but they’re not always sure how to describe it clearly or which adjustment actually matches what they’re feeling.
So I built a web app called Trailogic around that idea. You can save your bike setup, describe what the bike is doing on trail, and use it to get a clearer next change instead of guessing or changing a bunch of things at once. It also has a couple free tools for spring-rate baseline and quick tuning help.
I’d be interested in honest feedback from riders here, especially on:
- whether the tuning flow makes sense
- whether the recommendations feel useful or too generic
- whether this is something you’d actually use between rides
Link: https://trailogic.app
Not claiming it replaces real experience or a good suspension tuner. I built it to help riders get from “something feels off” to a better next step.
Curious why I wouldn’t just use the underlying model instead of an app. Seems the model by itself would be far superior.
Fair question. You probably could use the raw model on its own, but what I’m trying to build is more the structure around it.
For suspension tuning, the hard part usually isn’t just getting an answer, it’s getting consistent input, keeping the bike/setup context, and narrowing things down to one clear next change instead of a broad conversation that can drift around.
So the app is really meant to make that process more repeatable between rides, not just wrap a model for the sake of it. Still early though.
My take: a large context SOTA model with tool calling (web search + code sandbox) in a simple chat interface will probably outperform a purpose-built web app, unless you've got something heavily quantitative on the suspension side that actually needs a database.
Think about it this way: over a full summer of suspension work, you're highly unlikely to exceed the context window of a modern model. That means a chat thread can hold your entire setup history, and if you prompt it properly you'll get pretty dang good results without needing any custom infrastructure.
If you wanted to get clever, you could build a "skill" specifically for suspension tuning (basically a reusable prompt + reference doc the model loads on demand). I bet the Vital community could contribute to that in meaningful ways. And if you wanted tighter control over memory and more flexible tool calling, you could build it into OpenClaw.
You can see a screenshot of ~5 minutes of prompting here. With a skill I bet it’d kick butt.
That’s actually a good example of what I mean.
Once you move from a blank chat into a structured tool with saved setup, specific inputs, and a repeatable tuning flow, you’re already pretty close to what I’m trying to build. The value to me isn’t just “use a model,” it’s making the process more consistent and useful over time for riders.
A strong general model can definitely get good answers. I’m just betting that for suspension tuning, a purpose-built workflow with bike context, setup history, and more controlled input/output is more useful long term than starting fresh in chat each time.
And to be fair, what you built there looks cool.
Post a reply to: Looking for feedback on a suspension tuning app