In-depth review of Kotcha, the AI-powered run coaching app co-founded by the greatest marathon runner of all time, Eliud Kipchoge, and trained on his (and his team’s) methodology and expertise. Want to dive straight in? Here’s our link for a two week free trial of Kotcha.

AI-Powered Run Coaching

Whether you use one or not, you’re likely aware of the existence of AI-powered running coaching apps. In fact, there’s a good chance you’ve read one of the several reviews we’ve written about Runna — the category’s biggest player. Love it or hate it, AI-powered run coaching is here to stay. Strava’s acquisition of Runna proves as much, and signals to other business-savvy folks there’s potentially some serious cheddar to be made. While it’s far from a saturated space, there’s been a wave of new entrants recently. The question, therefore, is how do these new brands sufficiently differentiate themselves from the likes of Runna and capture market share?

Having Eliud Kipchoge, the greatest marathon runner of all time, as co-founder — and building the AI on his and his team’s methodology and expertise — is, it’s fair to say, a pretty strong differentiator.

That’s precisely what Kotcha have done, billing themselves as “Kipchoge’s running coaching app.” Seeing the Kenyan’s name as co-founder certainly caught my attention — but I was, just as you may be, cynical about the extent of his involvement. Was Kipchoge merely the face of yet another AI-powered training plan, or did his and his team’s approach to training inform the model, and to what extent?

To answer those questions — and to explore what really separates Kotcha from the likes of Runna — I spent the last ~12 weeks training with it. I also reached out to Dimitri Dor, one of Kotcha’s co-founders, who shed some light on how the app adapts to individual runners, the true extent of Kipchoge’s involvement, and where Kotcha is headed next.

How’s It Work?

Plan Generation

First, let’s talk about plan generation. There’s a series of onboarding questions to help the system learn about your experience, goals, and how often you’d like to train. They’re reasonably analogous to the ones used by Runna (and likely other run coaching apps) which makes sense. Really, there’s not much more you could ask before starting a plan. However, when it comes to plan generation, that’s where the similarities end. Where Runna takes the self-reported data and spits out a plan for the entirety of your selected duration, Kotcha says “now prove it.”

Before generating a plan, you’re required to connect a compatible smartwatch (there’s currently support for Garmin, Apple, Coros, and Huawei) and log a few runs. Then, every Sunday at 6PM, the next week’s training becomes available. You simply tap the screen to reveal it, log how you’re currently feeling, and hit validate to agree to the proposed workouts.

That’s because one of Kotcha’s big selling points is its adaptive engine. What’s revealed next week depends on your performance — and how that performance felt — in the present week.

As adaptation seems to be a word thrown around quite freely in this space, I wanted to find out exactly which elements of the plan were actually subject to change. I put the question to co-founder Dimitri:

The plan adapts your target paces, weekly volume, which workouts appear and some workout details (reps/durations). Adaptation exists to keep progression safe and realistic: the engine balances planned progression with real-world results and fatigue. Changes are conservative by design so the plan nudges you rather than swings wildly. 

A single short week is treated gently: the system smooths volume so one bad week usually won’t collapse the next long run. If you miss multiple weeks or repeatedly fail important sessions, fitness estimates fall and the plan reduces intensity/volume more noticeably. In short, small slips produce small corrections; repeated slips produce larger, systematic downgrades.

Plan Structure & Workout Variety

On the five-sessions-per-week plan, Kotcha’s training resembles something close to an 80/20 split: two easy runs, one long run typically performed at an easy pace, and a couple of faster-paced sessions — a mix of tempo, intervals, and target race pace workouts. Two of the five sessions are labelled “priority” (previously “must do”), flagging them as the week’s non-negotiables. In training for 42.2, the long run is almost always one of them. Beyond running, Kotcha recently added strength training sessions to the plan — exercises drawn directly from Kipchoge’s own routine, each accompanied by written descriptions of form and technique.

Post-Run Feelings Tool & AI Interaction

At the end of each run, you’re asked to rate how it felt on a scale of 1–10, with each value accompanied by a descriptor so you know exactly what you’re logging. You can also add detail — either by typing or talking — to give the AI more context about how the session went. Kotcha’s AI, Coach K, uses this information in its post-run analysis, responding to what you’ve logged and flagging anything worth noting.

Beyond post-run feedback, you can interact with “Coach K” — Kotcha’s AI — more broadly. Ask it for a breakdown of a specific week’s training, or — in one of Kotcha’s more recent updates — request changes to upcoming sessions. Not feeling 100%? You can ask Coach K to shorten a run or dial back the target pace before you head out. Too much volume in your current week? Tell Coach K and see what it suggests.

Coach Kipchoge

One really neat feature is the messages you receive from the G.O.A.T, Eliud Kipchoge. At various intervals throughout your plan — after your first day, first week, five weeks in, the day before your race, and upon completion — you’re rewarded with a pre-recorded message. It’s a clever reward mechanism that’s surprisingly motivating, and one I genuinely admire from a business perspective; gamification that keeps you engaged with your plan and, fundamentally, your Kotcha subscription. They’re not personalised, but it’s likely the closest I’ll ever get to a chat with the man himself — so I’m all for it.

But as I mentioned in the introduction, I had my doubts about the depth of Kipchoge’s involvement. Are these videos — along with the use of his name — the extent of it? Dimitri was unequivocal:

Eliud is a co-founder of Kotcha, not just a partner, or an ambassador.

When we went to Kenya, we spent hours with him and his entire team. That time went into building a knowledge database that powers our AI coaching, so when a user asks a question, the answer reflects how his team actually thinks about training. We also worked with him and his team directly to design the plans, which are adapted to every level and goal.

The videos are just the most visible part of it. The deeper integration is in the training methodology and the AI layer. 

Having gone in sceptical, that was enough to satisfy me. Far from being cosmetic, Kipchoge’s involvement is the foundation the whole app is built on. If further evidence were needed, Kotcha recently introduced strength training sessions drawn directly from his personal routine — and Dimitri tells me he’ll be adding more content directly in the near future.

Gritty Verdict

I’ll start by saying I’ve really enjoyed training with Kotcha. I had to cut the plan short due to a climbing injury — nothing running-related — but in the 10 weeks or so I used it, I progressed faster than I have with any comparable app. In the course of marathon training, I smashed my 10K PB and feel confident in my ability to achieve PBs across the board when I’m back to full fitness. That’s down to the structure and variety of Kotcha’s plans.

The intensity is perfectly calibrated — a well-balanced mix of easy, moderate, and hard sessions. Drop the ego, run your easy runs truly easy, and you’ll find yourself genuinely ready to push when it matters. The progression feels equally well-managed; mileage builds gradually rather than aggressively, which makes the plan feel sustainable over a long training block rather than something you’re just surviving week to week.

The variety keeps things fresh too. Across a typical block you’ll encounter tempo runs, endurance runs, interval sessions, race pace work, and long runs. What’s more, there’s a tonne of variation within those sessions (particularly the speed-focused ones) — it’s not just the same session recycled week after week. For anyone who’s found other plans repetitive, that variety makes a real difference over a long training cycle.

It’s hardly surprising that the running component is so well-designed. Kipchoge and his team’s knowledge base forms the foundation — and the quality of the sessions is exactly what you’d expect from that pedigree.

That said, there are a few elements worth examining more closely — some that I think Kotcha get really right, and others where I think there’s room to do better.

Adaptation: Does It Work?

Kotcha’s adaptive engine has genuinely impressed me at times. On a couple of occasions, poor route choice left me running something like 7km instead of a prescribed 10. Rather than ignoring it, Kotcha read that as a potential sign of accumulated fatigue and quietly reduced pace targets for the following week. It’s a small thing, but it speaks to the safeguards built into the system and a clear intent to minimise injury risk.

Said safeguards are scattered throughout. For example, after a three-week layoff, I asked Coach K to increase my pace targets for a specific run. It suggested a marginal increase but flagged that mileage had been low, keeping the recommendation conservative. Cautious, considered, exactly what you’d want.

Which makes what happened next all the more baffling. After logging just one run in three and a half weeks, I expected the reveal to surface a significantly scaled-back plan. Instead, it threw me straight into peak week — 79km from a standing start. The AI clearly recognised the low mileage; it said as much when I asked about pacing. Yet the plan didn’t reflect that at all.

I’ve always stressed in these reviews that runners should question what an app tells them rather than follow it blindly. I’m not going from 0 to 79km in a week. But for a less experienced runner — one who chose Kotcha precisely because it bills itself as adaptive — that discrepancy could be genuinely dangerous.

Perhaps this is an edge case the algorithm isn’t yet equipped to handle. If so, the fix seems straightforward: something similar to Runna’s approach, where the app recognises a prolonged drop in mileage and asks whether the user wants to reformulate or start fresh.

Despite this, the adaptive engine shows real promise — better than anything comparable I’ve tested. The cautious-by-design philosophy is exactly the right approach for an app that takes injury prevention seriously. Get the edge cases right and it has the potential to set a new standard.

Post-Run Feedback Tool & AI Interaction

I’m a huge proponent of post-workout feeling tools — I find them useful for spotting patterns in training and recovery that aren’t always obvious in the moment. Kotcha’s implementation is one of the better ones I’ve used. The 1–10 scale with descriptors removes ambiguity, and the ability to add context by voice or text means Coach K has something meaningful to work with.

That quality of input shows in the output. On a recent easy run where I flagged a mild respiratory issue and logged a higher-than-expected effort rating, Coach K’s response was considered and appropriate. It even accounted for this issue in the feedback it provided two or so runs after the fact, making it feel a tad more personalised and boosting its “real coach feel.” It’s AI, I know — but it’s noticeably better than the generic feedback served up by the likes of Garmin and Strava.

Recently, Kotcha added more functionality to these interactions too. As the screenshot below shows, you can now ask your AI coach to change certain aspects — like target pace or total duration — of select runs. For long and easy runs, for example, you’ll be able to increase or decrease pace targets. Others, like race pace specific practice, have a few more restrictions. While this is a great feature, it’s one you might only know about if go looking (or if you’re fortunate enough to have one of the co-founders tell you about it!) More on that in What’s Missing.

Must Dos / Priority Sessions

I like the concept of priority sessions, but I think the execution carries some risk — particularly for newer runners. Take a user with a busy week who can only fit in a couple of sessions. They’ll naturally gravitate toward the ones the app flags as most important. But in marathon training, jumping into progressively longer long runs without the supporting easy mileage underneath them is a reliable route to injury.

I don’t for a minute think Kotcha intends to imply the other sessions don’t matter. But we have to consider who these apps are built for — and for many users, the label does the thinking for them.

There’s also a broader issue with the concept of priority in this context. A human coach can factor in stress, poor sleep, accumulated fatigue, and adjust what constitutes a priority accordingly. What feels like a non-negotiable long run one week might sensibly become an easy 5K and an early night the next. That kind of nuanced, contextual judgement is exactly what an AI-powered app — even an adaptive one — struggles to replicate. Kotcha therefore needs to either better contextualise what priority actually means, or reconsider the label altogether.

Kotcha vs Runna: Which Should You Choose?

Over the last two years I’ve written three Runna reviews, so I have a solid basis for comparison. I won’t go into exhaustive detail here — that’s what those reviews are for — but a few key differences are worth breaking down.

Who’s Behind It

Both apps have strong teams. Runna was co-founded by Ben Parker, a certified running coach, and Dom Maskell, a former McKinsey consultant, and they’ve since assembled an impressive roster of coaches and athletes including British Olympian Steph Kessell, US Olympian Colleen Quigley, and Australian Olympian Genevieve Gregson. It’s a credible, knowledgeable operation. But Kotcha has Eliud Kipchoge as co-founder — and his coaching team’s knowledge forms the foundation of the AI. Beyond Kipchoge, the team also includes athletes from the dsm-firmenich running team, formerly the NN Running Team, one of professional running’s most decorated outfits. On pure pedigree, Kotcha wins this one.

The Run Training

Both are genuinely good, but I prefer Kotcha’s run training. If you’re choosing an app purely on the quality of the workouts — variety, structure, balance, safe progression — Kotcha is my pick. You’ll make gains with either, but I find Kotcha easier to stick to, and the results have reflected that.

Ecosystem

This one goes to Runna, and it’s not particularly close. Runna has a far broader feature set — accessory training, community features, and a wider range of plan types. Kotcha is adding features at a decent pace, but as things stand they’re some way behind the overall package Runna offers. If you want more than a running plan, Runna is the choice.

Experience Level

Runna’s manual intensity adjustment makes it more immediately customisable, which could suit runners who already have a strong sense of what they need. But Kotcha’s adaptive engine does that calibration automatically — and arguably more accurately — which makes it well-suited to runners who want the app to do the thinking for them. Both work across experience levels; it just depends whether you prefer control or automation.

Price

Runna costs £15.99 per month or £99.99 annually. Kotcha is marginally cheaper at £14.99 per month or £99.99 annually. There’s little in it. Whether Runna’s additional features justify the difference in monthly price depends entirely on whether you’ll actually use them — but for the running plan alone, you’re paying roughly the same for both.

What’s Missing / How Could It Be Improved?

• I’ve already spent some time discussing this, so won’t go into too much depth. Personally, I find the ability to preview upcoming weeks problematic. It feels counterintuitive to one of Kotcha’s best features — the reveal.
• I would love to see some sort of tutorial, guide, or walkthrough explaining what the AI is capable of. It’s not that it’s complicated to use. After all, the majority of people know how to use LLMs like Claude or ChatGPT. But I do think there’s a slight disconnect. While there are other AI-powered running coaching apps, they don’t integrate AI in the same way as Kotcha. So, I think there’s a real possibility that users won’t think to ask Coach K to change a workout. They won’t ask it for a summary of a week’s training. A brief guide on how the app works and everything you can do with it would solve that.
• Compared with something like Runna, there are fewer choices when it comes to plan generation. At the moment, it’s limited to training for an event (5-42K) or running your first 5 or 10K. This isn’t a criticism. I’m happy Kotcha is attempting to nail the basics before introducing more niche plans. It’s just worth mentioning in case you were looking for anything specific, like a ‘post-injury plan.’
• While they have recently added strength training sessions built from Kipchoge’s actual routine, at present they’re just lists of exercises with (good) descriptions of how to perform them. I’d like to see animations or videos utilised to really elevate this side of Kotcha’s plans.

Would I Recommend Kotcha — and Who To?

A few things need addressing, but none of them would stop me recommending Kotcha. The app offers a genuinely unique value proposition — the opportunity to train on plans built from the knowledge, experience, and expertise of Kipchoge and his team. And as I’ve outlined throughout this review, that pedigree shows. The structure, intensity, progression, and variety of both the overall plan and the individual sessions are the best I’ve encountered from an AI-powered coaching app.

In terms of who I’d recommend it to — honestly, most runners. Kotcha covers everything from “run your first 5K” through to marathon PB attempts, and the adaptive engine means the plan moulds to you rather than the other way around. The one gap is ultra distances — nothing beyond 42K currently — though given where Kotcha are in their development, that feels like a sensible place to focus for now.

What gives me particular confidence in recommending Kotcha is the trajectory. In the three months I’ve been testing the app, there’s been a steady stream of new features and improvements. Dimitri was understandably protective of the specifics of the algorithm — it’s core IP, after all — but was clear that it’s continuously being developed and refined. For an app still in its relative infancy, that momentum matters. The issues I’ve flagged feel like the growing pains of a product moving fast in the right direction, not fundamental flaws.

Kotcha launched with arguably the greatest marathon runner of all time as co-founder. The foundations are exceptional. Where they go from here should be very interesting indeed.

Kotcha Free Trial Discount Code

Just follow any of the Kotcha links dotted throughout (or hit the below button) and you’ll get a two week (versus the usual one) free trial automatically applied to your download.

Related Articles

• For comparison, why not check out our initial impressions of Runna, post-plan thoughts, or the latest post-Strava acquisition review.
• Fancy a new running watch? Our review of the Garmin Instinct 3 Solar is now live.
• Need a pair of carbon plated trainers to see you through the race at the end of your plan? See what we think to the new Puma Nitro Elite 4s, or check out the latest deals on last season’s race day shoes.