Intervals.icu is the richest free training analytics tool available. AthleteOS is the tool that tells you what to do with that data. If you already have a coach, Intervals.icu free wins easily. If you’re self-coached, read on.
The Verdict Up Front: AthleteOS vs Intervals.icu
Intervals.icu wins on analytics depth and price. It’s free, it integrates with 17+ devices, and it shows you CTL, ATL, TSB, power curves, and aerobic decoupling in a single dashboard used by 160,000+ athletes.
AthleteOS wins on coaching. It reads the same data daily and generates a plain-English training decision: what to do today, how hard, and why. Intervals.icu deliberately does not do this. That gap is the entire comparison.
Here it is in one line: Intervals.icu shows the map. AthleteOS tells you which road to take.
What Intervals.icu Actually Is (And What It’s Not)
Intervals.icu started as an analytics extension for Strava. Its creator, David Tinker, went full-time on it in September 2024. Today it has 160,000+ athletes and has analyzed over 111 million activities.
The free tier is genuinely free, with no trial period and no bait-and-switch. You get the Performance Management Chart (CTL/ATL/TSB), power duration curve, interval auto-detection, zone distribution, HRV logging from Garmin and Oura, and a workout builder. The optional Supporter tier at $4/month adds Strava full-history import, weather analysis, and an annual training plan builder.
What it deliberately does not do:
- Prescribe daily workouts based on your current fitness state
- Adapt your plan when you miss a week with illness
- Factor in last night’s HRV reading before deciding today’s session load
- Give plain-English coaching feedback on what went wrong or right
- Automate taper for a target race date
Intervals.icu is an analytics dashboard. It was built to show you everything. What you do with that data is entirely up to you.
This isn’t a criticism. It’s a design choice. But for a self-coached athlete, it’s a real gap.
The CTL/ATL/TSB Model: Powerful Math With a Ceiling
Both Intervals.icu and TrainingPeaks rest on the same foundation: the CTL/ATL/TSB model, originally derived from Banister’s 1975 impulse-response model. Dr. Andrew Coggan simplified it into exponentially weighted moving averages with a 42-day time constant for fitness (CTL) and a 7-day time constant for fatigue (ATL).
Translation: your fitness score (CTL) is roughly a 6-week rolling average of training load. Your fatigue score (ATL) is a 1-week average. The difference is your form score (TSB).
The model is useful. It’s also imprecise at the individual level. A 2006 study by Hellard et al. validated the Banister model on 9 elite swimmers and found that the fitness decay constant had a coefficient of variation above 30% between individuals. The correlation between fitness and fatigue gain factors was 0.91, meaning the parameters were nearly redundant. Overall model fit reached R² = 0.79, which sounds acceptable until you realize that 21% of performance variance is unexplained by the numbers on screen.
Translation: two athletes with identical CTL numbers may need completely different training prescriptions. The model gives you a useful signal. It doesn’t give you a coaching decision.
This is why 91% of elite programs monitor training load (Halson, 2014), but the same research flags the primary bottleneck: not data collection, but turning those numbers into actions. Collecting data without a decision layer doesn’t help — it just adds more to watch.
See our deeper breakdown in CTL, ATL, and TSB explained.
Does Intervals.icu Tell You What to Do Tomorrow?
No. This is the question most comparison articles skip, so let’s be direct.
Intervals.icu shows you data. It does not generate a training prescription. It does not know your race date. It does not look at your suppressed HRV from last night and say “go easy today.” It doesn’t send you a workout. It doesn’t flag that your TSB has been negative for 18 days straight.
You can see all of that in the dashboard. But interpreting it and deciding what to do requires either a human coach or a tool built to make that call.
Research backs why this matters. Studies on auto-regulated training vs fixed-load training found an effect size of 0.64 (p < 0.001) favoring adaptation — and in shorter 5–7 week blocks, that effect size climbed to 0.87 (Frontiers in Physiology, 2021). Plans that don’t adjust to your current state systematically underperform. Intervals.icu, without a coaching layer, is a fixed-load world.
And the stakes are real. Around 60% of elite women and 64% of elite men runners have been overtrained at some point in their career (Kreher and Schwartz, 2012). One of the earliest detectable signs is elevated perceived effort at a given workload — a signal that requires daily contextual tracking to catch. Intervals.icu logs your RPE if you enter it. But it doesn’t connect that RPE trend to a decision.
Feature Comparison: Intervals.icu vs TrainingPeaks vs AthleteOS
| Feature | Intervals.icu Free | Intervals.icu $4/mo | TrainingPeaks $19.95/mo | AthleteOS |
|---|---|---|---|---|
| CTL / ATL / TSB chart | Yes | Yes | Yes | Yes |
| Power curve analytics | Yes | Yes | No (needs WKO5 +$179) | Yes |
| Aerobic decoupling detection | Yes | Yes | No | Yes |
| Workout builder | Yes | Yes | Yes | Yes |
| Daily workout prescription | No | No | No | Yes |
| Plan adapts after missed sessions | No | No | No | Yes |
| HRV-based readiness adjustment | No | No | No | Yes |
| Plain-English coaching feedback | No | No | No | Yes |
| Multi-sport load balancing | Yes | Yes | Basic | Yes |
| AI coaching layer | No | No | No | Yes |
| Race taper automation | No | No | Manual | Yes |
| Setup complexity | High | High | Medium | Low |
TrainingPeaks and Intervals.icu share the same fundamental gap: neither prescribes workouts or adapts plans. TrainingPeaks was built to deliver a human coach’s plan to an athlete. If you don’t have a coach, you’re paying $19.95/month for a calendar with some charts.
Intervals.icu is honest about this. TrainingPeaks marketplace plans cost $30–$150 each and are static — they don’t adjust when life happens.
Pricing: What You’re Actually Paying For
| Platform | Monthly Cost | What You Get |
|---|---|---|
| Intervals.icu Free | $0 | Full analytics, no prescription |
| Intervals.icu Supporter | $4/mo | Analytics plus advanced planning tools, no prescription |
| TrainingPeaks Premium | $19.95/mo | Calendar delivery, basic CTL/ATL chart, no AI |
| TrainingPeaks + WKO5 | ~$35+/mo | Advanced analytics, still no AI, complex setup |
| AthleteOS | ~$19.95/mo | Integrated analytics plus daily AI coaching prescription |
At $4/month, Intervals.icu Supporter is almost unfairly cheap for what it gives you. But note: Strava full-history sync moved behind the Supporter paywall. If you assumed “free” meant fully-featured, that’s where friction appears.
AthleteOS competes at the TrainingPeaks price point. The difference is what that price buys you.
A Case Study: The Same Data, Two Different Decisions
Take a cyclist called Jake — 41, self-coached, training for a 70.3 in 14 weeks. On a Monday morning, his data shows CTL at 68, ATL at 82, TSB at -14. HRV is down 12% from his 7-day baseline. Saturday’s long ride RPE was 8 out of 10, but pace was normal.
In Intervals.icu, Jake sees those numbers. He knows TSB is negative. He knows HRV is low. He heads to the Intervals.icu community forums — and finds a thread where other athletes debate whether -14 TSB warrants rest or a light day. He trains hard anyway, because the plan says “threshold intervals.”
In AthleteOS, the AI coach reads the same inputs. It flags the suppressed HRV alongside the high RPE-to-load ratio from Saturday. It shifts Tuesday’s threshold session to an easy aerobic ride and pushes the interval work to Thursday, when form trends suggest readiness. Jake doesn’t have to interpret anything. He gets a sentence: “Your recovery metrics are below baseline. Today’s session is a 75-minute easy ride at Zone 2. Thursday is back on track for the threshold set.”
Eight weeks later, Jake’s TSB peaks at +6 on race morning — the kind of precision that requires daily adjustment, not a fixed plan.
HRV-guided training shows significant advantages for cardiac-vagal recovery (SMD+ = 0.50) and positive trends for VO2max and endurance performance (Manresa-Rocamora et al., 2021). The mechanism isn’t magic — it’s simply not ignoring signals your body is sending. See also how heart rate accuracy affects these readiness signals.
Which Platform Is Right for You?
| Your Situation | Best Fit | Why |
|---|---|---|
| You have a human coach already | Intervals.icu Free | Supplements coaching with deep analytics at $0 |
| Self-coached, want daily prescription | AthleteOS | Adaptive coaching without human coach fees |
| On TrainingPeaks, want deeper analytics | Intervals.icu Free | Free analytics layer alongside TP calendar |
| Budget-conscious triathlete needing structure | AthleteOS | Replaces $150+/month coach with AI-guided periodization |
| Beginner endurance athlete | AthleteOS | Intervals.icu’s manual setup is a barrier; AthleteOS configures automatically |
| Coach managing 20+ athletes | TrainingPeaks Coach Edition ($21.99/mo) | Built for multi-athlete delivery |
Where Intervals.icu Wins (Honestly)
Intervals.icu is better than AthleteOS in one specific situation: you’re already a data-literate athlete who knows how to read training load curves, and you want the deepest possible analytics at zero cost.
The power duration curve with multi-period overlays, the custom chart builder, W’ prime tracking, aerobic decoupling charts (learn more about what aerobic decoupling measures) — these are genuinely excellent. The platform has been iterated on by a single developer for years and it shows. The community is helpful. The integrations are comprehensive.
If you’re paying for a coach who sends you workouts, adding Intervals.icu free on top gives you analytics your coach probably doesn’t have time to provide. That’s a strong use case.
Where it doesn’t win: when you’re the person who has to decide what to do with all of that data. A 2025 BMC Research Notes study (n = 119 athletes) found that 25% of structured-plan users were using AI-generated plans — and those athletes reported significantly higher trust scores than athletes on human-designed plans (3.89 vs 3.31, p < 0.001). The appetite for AI coaching is real, and it’s growing.
Researchers have also flagged “paralysis by analysis” as a documented outcome when athletes face unsynthesized wearable data without an interpretation layer. More data without more clarity doesn’t help. It just creates more to worry about.
The AthleteOS vs Intervals.icu Decision in One Grid
Use Intervals.icu if:
- You already have a human coach sending you workouts
- You’re a data-savvy athlete who enjoys building custom charts
- Budget is the primary constraint and $0 is what you need
- You train primarily one sport and don’t need multi-discipline load balancing
Use AthleteOS if:
- You’re self-coached and need daily workout prescription, not just data
- You want your HRV, RPE, and load data synthesized into a single decision
- You’ve been on Intervals.icu for months but still don’t know what to do tomorrow
- You want your training to automatically adjust to illness, travel, or a hard week at work
Both platforms integrate with Garmin, Strava, WHOOP, and Oura — so starting with one doesn’t lock you out of the other. Many athletes run both: Intervals.icu for deep analysis, AthleteOS for daily coaching decisions.
Frequently Asked Questions
Is Intervals.icu actually free? Yes, with no trial limits. The $4/month Supporter tier adds Strava full-history import and advanced planning tools. Core analytics are permanently free.
Does Intervals.icu tell me what workout to do tomorrow? No. It shows your training load, power curve, and fitness trends. Training prescription is outside its scope by design.
What’s the difference between an analytics tool and an AI coaching platform? Analytics shows you numbers. A coaching platform reads those numbers and tells you what to do next. Intervals.icu is analytics. AthleteOS is coaching.
How do I know if my CTL/ATL/TSB numbers mean I should train harder or rest? That’s exactly the question neither Intervals.icu nor TrainingPeaks answers for you. The Banister model underlying both has a CTL time constant CV above 30% between individuals — meaning the same number has different implications for different athletes. Daily context (HRV, RPE, sleep) is required to interpret it correctly. AthleteOS synthesizes that context automatically.
Can I use Intervals.icu and AthleteOS together? Yes. Intervals.icu provides chart depth and custom analytics. AthleteOS provides daily prescription. They serve different needs and can complement each other if you want both.
Understanding Zone 2 and its role in aerobic base building helps make sense of why training load interpretation matters so much. If you want to understand how FTP anchors the TSS calculations both platforms use, what is a good FTP covers the benchmarks by athlete tier.
Ready to stop reading your data and start acting on it? AthleteOS connects to your existing Garmin, Strava, or WHOOP data and starts generating daily coaching decisions from day one — no manual zone entry, no setup spreadsheet. The analytics are already there. The direction is what’s been missing.