ElevenLabs expanded its MCP so supported assistants such as Claude, ChatGPT, and Cursor can call voice, transcription, dubbing, music, sound-effect, image, and video tools through one connection. OAuth handles sign-in, so users do not have to run their own MCP server or manage API keys. Results land in the ElevenCreative workspace, where they can be edited further in Studio. Read ElevenLabs’ announcement.
What one connection can do
In an assistant chat, users can request a voiceover, transcribe a recording with speaker labels and timestamps, dub a script into another language, create music and sound effects, or generate images and video. ElevenLabs says the connector draws on more than 50 models; available tools depend on the client and account permissions. Since outputs are stored in ElevenCreative, teams can manage assets in one workspace instead of repeatedly downloading and uploading them between tools.
Build a simple content workflow
- Add ElevenLabs MCP in a supported client and sign in with OAuth. Confirm the intended workspace and data region.
- Begin with one task, such as transcribing an interview with speakers and timestamps, then drafting a short-video voiceover.
- Specify language, audience, voice style, duration, and material that must not change. Obtain permission before dubbing while preserving a speaker’s voice.
- Review audio, text, images, and timing in ElevenCreative or Studio. Check rights, spelling, and commercial use before publishing.
Permissions, privacy, and cost
OAuth reduces manual key handling, but the connector can still act on the capabilities a user authorizes. Administrators can restrict available tools and select a data-residency region during setup. Apply least privilege and enable only what the current workflow needs. The announcement does not provide a universal free allowance or one price for all accounts; check current ElevenLabs plans and workspace settings.