From draft to audio in one chat
Ask your assistant to draft a 30-second intro, refine the wording together, and convert the final script into speech without copy-pasting.
Give your AI assistant a voice. Connect Voiceslab via MCP, ask in plain language for a voiceover, multi-speaker dialogue, or narration in your cloned voice, and get a download link right in chat.

Text to speech MCP
Voiceover workflow
Script to voiceover
Preview, refine, export
WORKING VOICE DEMO
The MCP server uses the exact same voices as this demo. Test a sentence here, then ask your assistant to narrate an entire script.
WHAT IT IS
MCP (Model Context Protocol) is an open standard that connects AI assistants to external tools. An MCP server exposes actions the assistant can take and lets it invoke them automatically whenever your prompt calls for it.
The Voiceslab MCP server gives your assistant five tools: list voices, convert text to speech, generate multi-speaker dialogue, retrieve a past generation, and check your credit balance. Because it runs remotely on our servers, there is nothing to install locally, and it uses the same credits as your Voiceslab plan.
SET UP IN A MINUTE
In Claude.ai, ChatGPT, and other apps with connectors, add a custom connector using https://voiceslab.io/api/mcp and sign in. In Claude Code, Cursor, and similar developer tools, add the server with your API key:
See setup for every clientclaude mcp add --transport http voiceslab https://voiceslab.io/api/mcp \
--header "Authorization: Bearer $VOICESLAB_API_KEY"WHY PEOPLE USE IT
Writing a script and producing the audio no longer have to happen in separate tools.
Ask your assistant to draft a 30-second intro, refine the wording together, and convert the final script into speech without copy-pasting.
Generate narration for product demos, tutorials, or in-app prompts directly from Cursor or Claude Code while you build.
Tell your assistant to use your cloned voice so every clip it generates sounds unmistakably like you.
HOW IT WORKS
Setup takes about a minute in most supported apps.
Speech generation via MCP is available on every paid plan and also works with credit packs.
Add https://voiceslab.io/api/mcp as a custom connector, or configure it in your client's MCP settings with an API key.
Sign in to Voiceslab and authorize the client to access your account. You can revoke access anytime.
Prompt your assistant naturally—for example: "Read this paragraph in a calm female voice and give me the MP3."
WHAT YOU GET
Designed for conversational AI assistants and AI coding editors alike.
Connect seamlessly with Claude, ChatGPT, Claude Code, Cursor, Windsurf, VS Code, and any client supporting remote MCP servers.
Chat apps with connectors use a secure Voiceslab sign-in, while developer tools can authenticate with an API key from your dashboard.
Your assistant can search the voice library by language, gender, or style, or use the private voices you've cloned.
Generate single-voice speech up to 11,000 characters per request in MP3, WAV, or Opus, or multi-speaker MP3 dialogue with up to 5 voices.
Every generation returns a download link valid for 1 hour, while the audio remains stored in your account for 7 days.
Each response reports the credits used and your remaining balance, and failed generations are never charged.

Direct the delivery
Your assistant can search the voice library by language, gender, or style, or use the private voices you've cloned.

Build for more audiences
Every generation returns a download link valid for 1 hour, while the audio remains stored in your account for 7 days.
GET BETTER RESULTS
Be specific about the voice you want. Mention language, gender, and tone—such as a warm British narrator—or name a specific voice you already like. Your assistant will search the library and pick a match, and can also adjust speed and volume.
Have the assistant prep your script before generating audio. Ask it to tighten long sentences, spell out numbers, and add punctuation for natural pauses. Before running long scripts, ask it to check your credit balance.
QUESTIONS BEFORE YOU CREATE
Any client that supports remote MCP servers, including Claude, ChatGPT, Claude Code, Cursor, Windsurf, and VS Code. Clients that only support local stdio servers can connect via mcp-remote.
No. The server is hosted at https://voiceslab.io/api/mcp. Simply add the URL to your app and sign in, or authenticate with an API key.
MCP requests use your Voiceslab credits just like the studio: 1 credit per character of input text in every language. Free accounts can connect, list voices, and check their balance, while generating speech requires a paid plan or a credit pack.
Yes. Any voice you clone in the Voiceslab studio appears in your assistant's voice list and remains private to your account.
Open Dashboard → API & MCP to disconnect the authorized app or revoke its API key. Access stops immediately.
Yes. Developers can access the same speech, dialogue, and voice capabilities directly through the Voiceslab text to speech API.
Connect Voiceslab in a minute and turn any script in chat into natural speech.
Connect Your AssistantText to speech MCP
Voice track