Elon Musk’s personal lawyer Alex Spiro sent a letter on behalf of Twitter to Microsoft CEO Satya Nadella that accuses Microsoft of violating Twitter’s developer agreement, as reported by The New York Times. In the Thursday letter (available in full below), Spiro claims that Microsoft may have been in violation of “multiple provisions” of Twitter’s developer agreement “for an extended period of time.”

Until recently, Microsoft had integrated Twitter’s APIs with some of its products — Twitter’s letter claims Microsoft operated “eight separate Twitter API apps” integrated with products like Xbox, Bing, and its ads platform. But it began pulling support for those APIs in April, including dropping clip sharing to Twitter from Xbox, likely due to Twitter’s new and generally more expensive API tiers.

According to the letter, examples of Microsoft violating Twitter’s developer agreement include using the API for “unauthorized uses and purposes” and Microsoft not informing Twitter of “any use case for six of the eight Microsoft Apps that it continued to operate until last month.” Claiming their agreement requires full cooperation from Microsoft, Spiro’s letter requests a “compliance audit” for each of the eight apps through April 2023, when Microsoft shut them down. It demands a broad swath of information from Microsoft about each of the apps and how it used Twitter data, all provided by June 7th.

Microsoft spokesperson Frank Shaw shared the following statement with The Verge. “Today we heard from a law firm representing Twitter with some questions about our previous use of the free Twitter API. We will review these questions and respond appropriately. We look forward to continuing our long term partnership with the company.” The company also told the NYT that it doesn’t currently pay Twitter for data.

Twitter’s press email replied to a request for comment with a poop emoji, as it has since March.

Spiro’s letter doesn’t mention any pending legal action. However, in April, Musk already threatened to sue Microsoft because it “trained illegally using Twitter data,” seemingly referring to data used to train large language models used in chatbots like Microsoft’s Bing and OpenAI’s ChatGPT. Other companies have taken less drastic approaches to try and benefit from training those models; Reddit, for example, recently announced API changes to let it better monetize its data.