Modulate analyses live voice chat for harmful behaviour, which is the part of online moderation that text tools cannot reach. Platforms have spent a decade building text moderation while voice channels, where a great deal of the worst behaviour happens, stayed effectively unmoderated because nobody could listen to all of it.
The distinction it draws matters: identifying genuine harm rather than flagging profanity, since a swear word between friends and targeted abuse sound nothing alike in context and a keyword filter cannot tell them apart.
Pricing is usage-based and published, from around $0.025 per minute. Voice analysis at scale is a serious privacy consideration that platforms need to disclose to users, and context judgements about speech will always produce contested calls at the margin.
