In recent years we’ve come to rely on a familiar suite of training tools: TrainingPeaks for planning, Zwift for executing the workouts, Strava for tracking rides, and the more specialist MyWindSock or Best Bike Split for time trial strategy.
But Stride, the new training platform co-founded by former pro Alex Dowsett, aims to bring everything together in one place: “By combining WorldTour‑level performance insights with accessible, everyday‑friendly planning tools, Stride bridges the gap between complex performance software and the realities of time‑crunched cyclists and amateur racers,” says the press release.
According to Dowsett: “Most riders are drowning in data but starving for insight. Stride pulls your numbers together, explains what they mean in plain language, and turns them into a clear, adaptable plan you can follow around work, family and real life.”
Perhaps slightly cheekily, Dowsett announced Stride from the Zwift stage at Rouleur Live in November 2025. We sat down the following month with Alex Barlow, the software engineer behind the platform, and Julian Roberts, Stride’s head of growth, to talk about how it all started, why exactly cycling is drowning in data but starving for insight, and why the team are so certain that Stride’s AI planner will transform the way we train.
The idea for Stride began with an impromptu reevaluation of the industry’s leading training platform. “I was approached by a very good software engineer also called Alex,” recalls Alex Dowsett. “I had never considered that TrainingPeaks could be improved on, but when Alex said in his opinion TrainingPeaks was good 10 years ago and hasn't really moved on, it did make me think about it."
Barlow had already started building software to replace his own training tools, but was looking for input at a pro-racing level to take it further. “Alex Dowsett said he was excited about replacing training software,” Barlow says, “but what really excited him was the stuff he needed to do his job [as a coach and as Astana's performance engineer] that just didn’t exist.”
Having spent 12 years racing at the highest level with 15 pro wins in his palmarès, Dowsett had never been short of data – it’s just that there wasn’t a way to get it to show him what he wanted to see. And in his current role with Astana he was even more frustrated that he couldn’t turn the riders’ numbers into meaningful insights.
“One of the moments where I realised how much potential there was,” he says, “was during the [2025] Vuelta team time trial. I had a bunch of files that represented one ride, but I had no way of viewing it as a whole. I’ve got these riders and I need to explain to them how to ride it. It’s hard to just use words – I want to use data too. So Alex created software to enable me to layer all the riders’ files on top of each other. You can bring everyone’s rides in line and press play. It’s very satisfying – you watch all the dots move as the race unfolds.”
He continues: “On a fundamental level Stride is here to try and replace TrainingPeaks, but also to replace all the other platforms that teams and riders use. MyWindSock, Best Bike Split, elements of Strava and some other stuff on top of that.”
One of Stride’s most compelling features for Dowsett evolved out of a thought he’d had during his pro career as a time trial specialist. “I always thought it would be fascinating to see myself like in a PlayStation game – getting a ghost version of your fastest lap to chase,” he says, “not only against my fastest time around the club time trial, but also against Filippo Ganna, where is he pulling away from me? Or where am I holding him? So you would select your circle of who you want to compare yourself against, your circle, clubmates, it might be the entire race, and then you can bring everyone’s rides in line with each other.
“Strava does this but it’s difficult, you have to create your own segments, but Stride can now break down every single corner in a race.” He continues: “We can’t do racing lines, but we can show average speed, power through the corner, minimum speed, and how much time you gained or lost.”
Barlow continues: “Some riders aren’t huge watt monsters, but they’re incredibly smooth and make up time through bike handling. In a general use case you’d have two athletes, one guy has got loads more power than the other, but he’s not as fast. So the question is whether that’s down to aerodynamics, bike handling skills, the course, the weather… and our tooling helps you start to narrow that down. I was amazed how much time is lost in corners. With some of the big watts guys, if you just took their cornering up, they’d be next level Pogis or Remcos.”
Barlow says he was surprised by how much value his analysis tools could add to WorldTour level setups. “A lot of big teams have data guys,” he says. They’re not software engineers, they know how to get something from the data, but then the presentation and the tooling isn’t their forte. So I built Alex a couple of random tools, gave it to him and his response was, ‘you don’t know how mindblowing this is because we just don’t have this in the WorldTour’. So that’s how it started helping me do my job.”
Training tactics
Another key aspect of Stride’s development focuses on something that’s overlooked in most other training software: racecraft. “A lot of platforms focus on physiology,” says Dowsett, “but there’s very little about how you actually race.” That’s where Stride’s ‘matches’ feature comes in.
Julian Roberts explains how this works through his own recent race analysis. “After the race you can see exactly where you burned matches. We count a match as when you go into zone five and above.” Stride maps these high-intensity, energy-costly efforts onto all the graphs of the ride it creates; its AI pulls out everything that’s over a rider’s threshold. “Alex Dowsett described it once as being like having £100 to spend but you’ve got to decide where to spend it.”
In Roberts’s race, Stride’s analysis revealed a major tactical error. “I burnt two matches quite early on,” he says. “I was trying to chase onto what I thought was the break. I burnt 603 watts over 24 seconds but the group behind were closing in so I sat up. I should have checked how everyone was reacting behind me before I decided to go. I could digest all this afterwards.”
This level of post-race analysis can be invaluable, demonstrating again that watt monsters don’t always win. “Training is one part of it,” says Barlow. “How you race is the other.”
While the racecraft tools might be a revelation for racers and WorldTour teams, its biggest advantage for the “everyday athlete” is its AI-powered system, explains Barlow. When a rider signs up to Stride, all the rides they’ve ever uploaded to other platforms – once permission is given to access them – are analysed and metrics such as FTP, CP (critical power) and W’ (W prime) are automatically calculated. Stride’s AI also continuously rechecks these metrics: “If you demonstrate that you’ve got more than that, we recalculate everything. Something we do that most platforms don’t, we estimate FTP but then we go back and recheck your FTP against the rides. For example some systems will do this, then you’ll go and do an hour ride and get 110 TSS [training stress score]. We know that’s not physically possible, so we’ll go back and back… Stride does that until it’s happy with what the training output is.”
Then there’s Stride’s AI training planner which is, in Barlow’s words "really unique to Stride”. He explains: “With the traditional approach, you’ve got this layout in your calendar and it’s happening. It doesn’t matter what else is going on… if you have a VO2max session in the books it’s getting done. If you’ve got four hours on Christmas day morning, it’s happening. But normal people don’t work like that. They work late, they do different things, they have three beers instead of the one they thought they would have. Our system will adapt around that. One of the reasons why people randomly Zwift is because they think there’s no point in planning two weeks ahead because they’re not going to hit it.”
Roberts continues: “You have a conversation with it… you can go into a workout, edit it and just say ‘make this easier’ and it will take the whole thing down. Or remove one of the intervals… you can basically do anything you want with it. You could say something like ‘road race simulation with a climb in the middle and a sprint at the end’, and it will generate the workout for you. Leave your RPE and comments and the AI will learn from what you said, it will match it with your sleep data and perhaps will decide OK, we’re not doing a VO2 max session tomorrow.”
Barlow continues: “Essentially what you can do is comment, leave notes, illness, availability, if you’re going skiing. So you ask it to schedule next week and the next, and it takes the entire context of everything you’ve done, the history, previous workouts, and it will come back and generate a plan. When you come back from holiday it will have a complete overview. It will learn from you. It has all the history and conversations, who you are. It’s a bit like ChatGPT – it’s not ChatGPT, but it does have the whole context of mankind and training, so you don’t have to be simple with it.” He concludes: “The AI chat is the first really big thing that will be a win. Once you’ve used it, you will find it very difficult to go back.”
So who is the “everyday athlete” exactly, the user that Stride is targeting? Barlow says: “They might be someone training for a big event like the Étape du Tour, or someone who wants to crush their local time trial series or just get faster on group rides.”
These riders often fall into exactly the gap Stride says in its press release that it aims to bridge. “Most coaches charge a minimum of £250 per month and you're not even getting one on one,” says Roberts. “The coach will give them a plan, go in once a month, do a quick overview. It’s not personalised. We’re not here to replace coaches but we’re here to bridge the gap for people who do not want to pay for a coach but want to hit their cycling goals. The AI tools are there for those people.”
Stride costs £12 per month or £120 a year and offers a 30-day free trial at the time of writing.
Barlow adds: “I still think there’s a place for coaching in the world. Cyclists are bad at self evaluation. You can lie to ChatGPT the same way you can lie to anyone else, but you can’t lie to a coach because they’ll see it, and coaches are good at making you do it. In the future we could do some interesting things though Stride with an option where for an extra payment you could have an evaluation by a real coach as a one-off every so often. That could be good for people who just need a little bit of reassurance or have questions.”
Looking to the future, the founders see Stride as part of a wider shift in how cyclists train. Barlow points to the impact of platforms like Zwift, which he says has been “the biggest revolution I’ve seen. A turbo trainer was really basic and you were just a masochist if you used it. Most people didn’t have a power meter 15 years ago; people’s fitness was all over the place. Zwift has really brought people into understanding the numbers and that has been huge for the industry, it changes the way people think about cycling.
Roberts concludes: “We’re not here to aggressively go to users of TrainingPeaks and say ‘you should just come to Stride’ but we’ve built the tools, we’re offering it in a more manageable way, we’re aiming it at everyday people as well as elite cyclists. We’re here to guide them on a better journey.”
For all the information and to sign up visit Stride’s website, and keep an eye out for our review after training with Stride for six weeks.
