Strava Stays Strong Against Scrapers Ahead of IPO

AI firms have transitioned into organizations primarily focused on data, resulting in a heightened demand for increasingly large datasets for their training needs. In response to this demand, many AI startups are disregarding crucial internet protocols—such as adhering to robots.txt files that indicate which sections of a website should not be accessed by automated scrapers—opting instead for more intrusive data extraction techniques. Consequently, numerous websites have begun enforcing access limitations, with some even engaging in licensing negotiations with AI companies. Strava, a service aimed at fitness enthusiasts and social runners, is adopting such measures by restricting access to its site and implementing fees for developer use.

To address scraping concerns, Strava is bolstering its platform’s security and will restrict specific data access to authenticated users only. Previously accessible public profiles and fitness club listings will now require user authentication to view, thereby minimizing unauthorized scraping by AI.

In terms of the API, developers once enjoyed a free, tiered access system on Strava’s platform—starting with basic access and increasing as their projects progressed. However, the company is now implementing a standard fee of $11.99 per month for all developers, with possible adjustments based on regional pricing.

Strava has reported a growth in its developer community, rising from 185,000 last year to 241,000 this year, and is committed to supporting them. In line with this, Strava intends to promote the Model Context Protocol (MCP), a new framework designed to allow AI assistants and applications to consistently access external data, granting Strava more control over data sharing.

Furthermore, the company plans to deactivate certain API endpoints—specific pathways through which external applications can access designated data, like club details—to protect user data. In 2024, Strava tightened its API guidelines, reducing its availability for AI training and imposing restrictions on third-party applications that display user data. These adjustments sparked backlash from developers who raised concerns regarding their applications’ potential decline.

While some developers may adapt to the subscription model, the discontinuation of certain API endpoints could still negatively impact applications that depend on them. Strava is allowing a 90-day grace period for developers before these changes take effect.

In a discussion with TechCrunch, Strava’s CEO, Michael Martin, expressed apprehensions that unregulated AI scraping could threaten the future of the public internet.

“AI companies are voraciously scraping public websites to satisfy their relentless quest for training data, which universally decreases site performance,” Martin stated. “We have recently observed multiple instances of significant performance deterioration. Besides scraping public sites, they are also trying to access our data through our API, flouting our API terms.”

He admitted that Strava has turned down data licensing requests from prominent AI labs. Martin specifically mentioned Perplexity, noting that the AI search startup attempted to mask its scraping activities via aggregator services to hide its source, despite being denied permission. This reflects Perplexity’s previous dealings in comparable situations.

Martin also underscored concerns regarding server overloads caused by inefficient API requests from collaborative applications, which often place undue strain on Strava’s infrastructure. This scenario resembles Meta’s justification for banning third-party chatbots from WhatsApp last year, citing system overload challenges.

The timing of these initiatives appears deliberate. Strava confidentially filed for an IPO earlier this year, and its data protection measures may aim to demonstrate responsible data stewardship to prospective investors. Martin quickly acknowledged similarities to Reddit’s API access limitations from 2024, clarifying that, unlike Reddit—which based API access fees on call volume, rendering it unaffordable for many developers—Strava’s flat fee structure is designed to nurture a thriving developer ecosystem.

“We want our users to feel empowered regarding their data and to trust us in its management and protection. At the same time, we are committed to ensuring that developers can continue to prosper and grow,” Martin concluded.