The United States Supreme Court has failed to grapple with the unique interpretive difficulties presented by social media threats cases. Social media make hateful and threatening speech more common but also magnify the potential for a speaker's innocent words to be misunderstood. People speak differently on different social media platforms, and architectural features of platforms, such as character limits, affect the meaning of speech. The same is true of other contextual clues unique to social media, such as gifs, hashtags, and emojis. Only by understanding social media contexts can legal decision-makers avoid overcriminalization of speech protected by the First Amendment. This Article therefore advocates creation of a procedural mechanism for raising a context defense to a threats prosecution prior to trial. Comparable privileges protect defamation defendants from having their opinions misconstrued as defamatory and allow them to have their liability resolved at an early stage of litigation, often avoiding the anxiety and expense of trial. This Article contends that criminal defendants in threats cases should have a similar defense that permits them to produce contextual evidence relevant to the interpretation of alleged threats for consideration by a judge at a pretrial hearing. In cases that cannot be resolved before trial, the context defense would entitle a defendant to produce contextual evidence at trial and have the jury instructed regarding the role of context in separating threats from protected speech. Although adoption of the context defense would be especially helpful in correctly resolving social media cases, its use in all threats cases would provide an important safeguard against erroneous convictions of speech protected by the First Amendment.