MyBikeFitting: Free AI Bike Fitting—Saves $200, but is it Trustworthy?

2026-02-14 | ProductHunt | Official Site

A rider on a trainer with AI real-time labels: Back 67°, Arm 143°, Hip 61°, Knee 134°. This is the core of MyBikeFitting—using a camera to replace a $200+ professional fitting.

30-Second Quick Judgment

What is it?: Open the webpage, ride in front of your camera (or upload a video/photo), and the AI automatically measures your knee, hip, and back angles. It then tells you whether to raise or lower your saddle and how to adjust your handlebars. The whole process takes 5 minutes, is free, and your data never leaves your device.

Is it worth it?: It's worth a shot, but don't treat it as the final word. Think of it as the "AI first-aid" of bike fitting—it helps you catch obvious problems (like a saddle that's too low causing knee pain), but for complex issues, you still need a human. For cyclists who have never had a fitting, this is the perfect zero-cost first step.

Three Questions That Matter

Is it for me?

Who is the target user?:

• Amateur cyclists who ride often but have never had a professional fit.
• People experiencing knee, back, or neck pain after riding but don't want to spend $200 at a shop.
• New or used bike buyers looking for a quick "ballpark" setup.
• People with home trainers (the results are best in a controlled environment).

Am I the target?: If you feel uncomfortable anywhere while riding, or if you finish a ride feeling "off" but can't explain why—you are the target user.

When would I use it?:

• Getting a new bike -> Use this for a baseline setup to avoid guessing.
• Knee/back discomfort -> Use this to check if it's a fitting issue.
• Changing saddle/handlebars -> Re-measure to see if your angles are still in a healthy range.
• Curiosity -> Satisfy your curiosity about your posture for free.
• Existing injuries or complex needs -> This isn't enough; you should see an IBFI-certified fitter.

Is it useful?

ROI Judgment: For someone who has never had a fit, it's absolutely worth 5 minutes. The worst-case scenario is losing 5 minutes of your time. If you have persistent pain, treat this as a "free screening" rather than the final answer.

Is it satisfying?

The "Aha!" moments:

• Zero Friction: No registration, no payment, no app download. Just open the site and go.
• Privacy Friendly: 100% of the data runs on your device. Nothing is uploaded to a server—a rarity for AI products.
• Instant Feedback: See your joint angles in seconds with specific suggestions like "Saddle +2cm."

By the numbers:

The website shows over 1,757 analyses completed and 274 reviews with an average of 4.7/5 stars—indicating a solid user base even before its recent ProductHunt launch.

The interface is clean and simple. Results are clear: Knee extension 32°, Hip angle 38°, Back angle 47°, Arm angle 162°, with a direct suggestion to raise the saddle by 2cm.

Real feedback from the AI bike fitting world (from competitors like MyVeloFit, representing the general experience):

"Since that AI bike fit, my knees have been literally pain-free! Seriously, zero pain." -- Bikes and Bacon

"Perhaps the most glaring shortcoming of an AI-powered bike fit is that there's no back-and-forth, no bike fitter asking 'How does that work for ya?'" -- Cycling Weekly

For Developers

Tech Stack

• Frontend: Browser-based web app (HTML/CSS/JavaScript).
• Backend: None! 100% client-side processing.
• AI/Models: Likely TensorFlow.js + MoveNet (or MediaPipe BlazePose)—browser-based pose detection models identifying 17 keypoints with 30+ FPS real-time inference.
• Core Logic: Keypoint detection -> Calculate joint angles (knee, hip, back, arm) -> Compare with optimal ranges from scientific literature -> Output suggestions.
• Data Input: WebRTC camera / Video file upload / Photo upload.

Core Implementation

Simply put: It uses browser-based pose detection to find your joints (hips, knees, ankles, etc.), calculates the angles, and compares them to the optimal ranges recommended by the Holmes method (e.g., the knee should have 25-35 degrees of flexion at the bottom dead center). If you're out of range, it tells you how to adjust.

Before the analysis, there's a questionnaire (bike type, pain points, goals, body type) to make the recommendations personalized rather than generic. The site cites academic papers like Millour G (2020), Bini RR (2011), and Ferrer-Roca V (2014) as the scientific basis for its angle ranges.

Open Source Status

• MyBikeFitting itself: Not open source.
• Similar Open Source Projects:
  • datarootsio/bikefitting -- TensorFlow.js + MoveNet Thunder, analyzes 10s video for knee angles.
  • luckyzee10/ai-bike-fitting-app -- MediaPipe real-time detection, client-side processing.
  • Kinovea -- Free open-source video analysis tool for manual angle measurement.
• Difficulty to Build: Low-Medium. The core tech (MoveNet/MediaPipe) is open source. The real barriers are: (1) Biomechanical knowledge of angle ranges, (2) Questionnaire logic and personalization, (3) Productization and UX polish. Estimated 1-2 person-months for an MVP.

Business Model

• Monetization: Currently completely free with no paid tiers.
• Why it can be free: No servers = no operational costs. The only cost is frontend hosting (virtually zero).
• Future Potential: Paid advanced analytics, brand partnerships (bike manufacturer referrals), or a Pro version with history tracking.
• Concerns: No business model means a potential lack of long-term development motivation unless the founder has other incentives (e.g., a portfolio project or lead gen for another business).

Giant Risk

Low risk of being crushed by tech giants because it's a hyper-niche tool. However, if Apple Health or Google Fit adds posture analysis, it could indirectly cover this. The real competition is industry-specific: MyVeloFit is already funded and has 100k+ fittings worth of data.

For Product Managers

Pain Point Analysis

The Problem: Cycling discomfort (knee/back pain) is common, but professional fitting is too expensive and inconvenient.

How painful is it?: Very. The data says:

• 14.8%-33% of long-distance riders experience knee pain (Source).
• 55.1% of amateur riders experience lower back pain in their lifetime (Source).
• Improper bike fit is a leading cause; one study found improper saddle height significantly correlates with knee pain (p<0.04).

Frequency: Medium. Done when getting a new bike, changing parts, or when discomfort arises. Not a daily tool.

User Personas

Feature Breakdown

Competitor Comparison

Key Takeaways

1. "Zero Friction" Strategy: No registration, no payment, no data upload. This lowers the barrier to entry to the absolute minimum—a great lesson for any utility product.
2. Personalization via Pre-survey: Understanding the user context before AI analysis makes the same angle data yield different, more relevant advice.
3. Scientific Credibility: Citing specific academic papers builds trust more effectively than just saying "based on AI algorithms."

For Tech Bloggers

Founder Story

• Founders: Names not disclosed.
• Background: A team of cycling enthusiasts and biomechanics experts.
• The "Why": The founder struggled with knee pain for months. Professional fitters were expensive and hard to book. After weeks of trial and error on YouTube, they thought: "Why not let AI do this?"
• Comparison: MyVeloFit's founder, Jesse Jarjour, is an IBFI Level 2 certified fitter with 10 years of experience. MyBikeFitting feels more like a "personal pain-driven indie project."

Controversies / Discussion Angles

• Angle 1: "Is AI fitting actually reliable?": A long-standing debate in the cycling community. Proponents say AI results are nearly identical to human fitters (per Cycling Weekly tests); opponents say fitting is subjective and five fitters will give you five different results.
• Angle 2: "How does free actually make money?": The logic of zero server costs makes sense for now, but what about the long term? No revenue often means no updates. Is this a passion project, a lead magnet, or waiting for a user base to monetize?
• Angle 3: "Privacy vs. Accuracy": MyBikeFitting is 100% local. MyVeloFit processes on the server, which theoretically allows for larger, more accurate models. It's a classic trade-off.

Hype Data

• PH Ranking: 3 votes (very low, just launched).
• Twitter/X Discussion: 4 tweets, all automated shares, 0 interaction.
• Website Data: 1,757 analyses, 274 reviews, 4.7/5 stars.
• Search Trends: Virtually non-existent. Google searches for "MyBikeFitting" yield almost no third-party content yet.

Content Suggestions

• The "Hook": "I replaced a $200 bike fitting with free AI—here's what happened." Personal experience + comparison tests always perform well in cycling circles.
• Trend Catching: AI applied to specific vertical niches. "AI bike fitting" is a great case study for the practical application of pose detection.

For Early Adopters

Pricing Analysis

No paid version. Completely free. No registration required.

Getting Started Guide

• Time to Setup: 5 minutes.
• Learning Curve: Low.
• Steps:
  1. Go to mybikefitting.com.
  2. Fill out the short survey (bike type, pain, goals).
  3. Choose input: Live webcam / Upload video / Upload photo.
  4. Film from the right side (drive-side), ideally on a trainer.
  5. Wear tight clothing and ensure even lighting.
  6. View your angles and adjustment tips.
  7. Adjust your bike and see how it feels.

Pitfalls & Complaints

1. No Interactive Feedback: The AI says "Saddle +2cm," but it doesn't ask how that felt. A human fitter would ask for subjective feedback and fine-tune accordingly.
2. Strict Filming Requirements: You really need a trainer, a side-on view, and good light. Without a trainer, accuracy drops significantly.
3. Limited for Injuries: If you have specific medical issues or body asymmetries, the AI might give generic advice that doesn't suit you.
4. Very New Product: Lack of extensive third-party validation for its accuracy.

Security & Privacy

• Data Storage: 100% local; no data is sent to a server.
• Registration: None. No email, no credit card.
• Privacy Policy: "No data sent to any server"—this is true privacy.
• Vs. Competitors: MyVeloFit and Apiir require video uploads; MyBikeFitting wins on privacy.

Alternatives

For Investors

Market Analysis

• Bicycle Market: $127.81B in 2026 -> $291.90B in 2034, CAGR 10.87% (Fortune Business Insights).
• High-end Bicycle Market: Growing by $6.5B (2025-2029), CAGR 7.5% (Technavio).
• Bike Fitting Market: No independent report, but with a massive global cycling population and traditional fits costing $100-$300, the room for digital disruption is huge.
• Pain Point Validation: 15-33% of riders have knee pain; bike fitting is the most recommended solution.

Competitive Landscape

Timing Analysis

• Why Now?: (1) Browser-based AI (MoveNet/MediaPipe) is finally mature enough for real-time mobile use. (2) Post-pandemic home training habits have stuck. (3) The rise of DTC bike brands means more people buy bikes without an in-store fitting.
• Tech Maturity: High. TensorFlow.js + MoveNet is stable for 30+ FPS inference.
• Market Readiness: Medium-High. MyVeloFit has proven the demand, but the "free" market is still up for grabs.

Team Background

• Founders: Anonymous; self-described as cycling enthusiasts + biomechanics experts.
• Team Size: Unknown.
• Track Record: Unknown—this is the biggest red flag for institutional investors.

Funding Status

• Current: No public info.
• Comparison: MyVeloFit raised a seed round led by former Shopify CTO Cody Fauser.
• Verdict: MyBikeFitting currently looks like an indie developer's side project rather than a VC-backed startup. Without a clear business model or transparent team, it's not an investment target yet, but it's worth watching if they hit 100k+ users.

Conclusion

In a nutshell: MyBikeFitting is the "AI first-aid" for bike fitting—technically mature (browser pose detection), smooth experience (5 mins, no registration), and excellent privacy (100% local). However, the product is very new, the team is anonymous, and the business model is non-existent, making its long-term survival uncertain.

Resource Links

2026-02-14 | Trend-Tracker v7.3