ElevenLabs releases a Voice Changer Skill

ElevenLabs releases a Voice Changer Skill for AI coding agents like Cursor, Claude Code, and Windsurf to integrate its Speech-to-Speech audio technology.

ElevenLabs Developers has released a Voice Changer Skill, a module designed for AI coding agents such as Cursor, Claude Code, or Windsurf. The "skill" format refers to a block of instructions, code examples, and best practices that the agent consults to produce functional code from the first iteration, without the user needing to manually consult the documentation.

The module provides access to ElevenLabs' Speech-to-Speech functionality, which converts a source audio file into a target voice (from the library, cloned, or created on the fly) while preserving the original performance, timing, emotion, intonations, and paralinguistic elements such as laughter or whispers. Two models are recommended: `elevenmultilingualstsv2`, covering 29 languages, and `elevenenglishstsv2` for English. Files are limited to five minutes or fifty megabytes, and the cost is billed at approximately one thousand characters per minute of processed audio. Target Voice ID, background noise suppression, output format, latency, and seed are among the options exposed by the skill.

Installation is done via a single command, `npx skills add elevenlabs/skills`, which places the module in the project's dedicated folder (`.cursors/skills`, `.claude/skills`, etc.) and simultaneously adds the other official skills from the publisher: Text-to-Speech, Speech-to-Text, Sound Effects, Music, Voice Isolator, and Agents. The source code is available at github.com/elevenlabs/skills, and the engine remains testable without integration from the playground elevenlabs.io/voice-changer.