Since Twitter launched in 2006, the company has acted as a kind of heartbeat for social media conversation. That’s partly because it’s where media people go to talk about the media, but also because it’s been willing to open up its backend to researchers. Academics have used free access to Twitter’s API, or application programming interface, in order to access data on the kinds of conversations occurring on the platform, which helps them understand what the online world is talking about. Twitter’s API is used by vast numbers of researchers. Since 2020, there have been more than 17,500 academic papers based on the platform’s data, giving strength to the argument that Twitter owner Elon Musk has long claimed, that the platform is the “de facto town square.” But new charges, included in documentation seen by WIRED, suggest that most organizations that have relied on API access to conduct research will now be priced out of using Twitter. It’s the end of a long, convoluted process. On February 2, Musk announced API access would go behind a paywall in a week. (Those producing “good” content would be exempted.) A week later, he delayed the decision to February 13. Unsurprisingly, that deadline also slipped by, as Twitter suffered a catastrophic outage. The company is now offering three levels of Enterprise Packages to its developer platform, according to a document sent by a Twitter rep to would-be academic customers in early March and passed on to WIRED. The cheapest, Small Package, gives access to 50 million tweets for $42,000 a month. Higher tiers give researchers or businesses access to larger volumes of tweets—100 million and 200 million tweets respectively—and cost $125,000 and $210,000 a month. WIRED confirmed the figures with other existing free API users, who have received emails saying that the new pricing plans will take effect within months.  “I don’t know if there’s an academic on the planet who could afford $42,000 a month for Twitter,” says Jeremy Blackburn, assistant professor at Binghamton University in New York and a member of the iDRAMA Lab, which analyzes hate speech on social media—including on Twitter. Twitter did not respond to a request for comment. For subscribers to the cheapest package, the number of rules through which they can filter data from the app’s Real Time PowerTrack API will be capped at 25,000, and the number of queries of the Full Archive Search API will be capped at 50,000. The number of Twitter handles they can analyze through the Account Activity API will also be limited to 5,000, and there will be a max of 20 requests per minute for the Engagement API Totals Endpoint, which allows researchers to see how well tweets are doing in terms of engagement. While this sounds like a substantial dataset, it only accounts for around 0.3 percent of Twitter’s monthly output, meaning it is far from being a comprehensive snapshot of activity on the platform. Twitter’s free API access gave researchers access to 1 percent of all tweets. Elissa M. Redmiles, a faculty member at the Max Planck Institute for Software Systems in Germany, says the new prices are eye-watering. “It’s probably outside of any academic budget I’ve ever heard of,” she says, adding that the price would put off any long-term analysis of user sentiment. “One month of Twitter data isn’t really going to work for the purposes people have,” she says. Kenneth Joseph, assistant professor at the University of Buffalo and one of the authors of a recent paper analyzing a day in the life of Twitter, says the new pricing effectively kills his career. “$42,000 is not something I can pay for a single month in any reasonable way,” says. “It totally destroys any opportunity to engage in research in this space, which I’ve in many respects built a career on.” The pricing documents were provided to WIRED by a researcher who asked for anonymity, since they are still accessing Twitter data through an existing API agreement and worry it could be terminated if they were identified. They say the new costs were “not viable for the academic community.” “No one can afford to pay that,” they say. “Even rich institutions can’t afford to pay half a million a year for a thimbleful of data.” It’s not clear whom the new pricing model is targeted at. Nir Grinberg, an assistant professor in the ​Department of Software and Information Systems Engineering at Ben-Gurion University in Israel, used to work at a startup that used Twitter’s data. “It seems like a really steep increase for a tiny amount of data. One percent of Twitter a few months ago was free. Now Twitter is offering 0.3 percent for half a million dollars [a year],” he says. “It’s just crazy. I honestly don’t know who could budget for this.” Researchers say the damage won’t just be to academic discourse. Twitter is a vital dataset for understanding how the internet works and what conversations are being had in the notional global public square. Joseph recognizes that there are other platforms he could research, but notes that Twitter’s potent combination of journalists, high-ranking politicians, and business decisionmakers makes it a vital area for research. “Twitter is a particularly special space for understanding elite discourse,” he says. “To rip that away from all of us trying to use the system to understand it is a tough pill to swallow.” Blackburn, however, says researchers will continue to find a way to scrutinize what’s happening on Twitter. “We’ve been mostly cut off from Facebook for years and we’ve continued to make progress,” he says. “It’s not like science is going to be held hostage by a guy that played himself into burning $44 billion on a website that makes no money, just so he could force all its users to read his shitposts.” You Might Also Like … In your inbox: Will Knight's AI Lab explores advances in AI