Voicely + n8n · workflow automation

Text to speech for n8n workflows

Turn workflow text into multilingual audio with the Voicely-maintained n8n community node. Choose a voice, generate speech over HTTP or WebSocket, and pass the audio file straight to your next step.

Get a Voicely API keyInstall the n8n node

Uses your paid Voicely API credits. Manage funding and daily API spend from your account.

Your text becomes a workflow asset

  1. Input

    Text from your workflow

    A script, article or text prepared by another node

  2. Voicely

    Generate speech

    Voice + language + HTTP or WebSocket

  3. Output

    speech.mp3

    Binary audio ready to store, send or publish

An n8n TTS node with HTTP, WebSocket and voice discovery

HTTP

Generate speech

Turn a script or a previous node's text into a complete audio file. Choose a voice, model, language and output format inside the node.

Audio binary for the next workflow step

WebSocket

Generate speech via WebSocket

Receive generated audio over a WebSocket connection. The node collects the audio and returns a complete file when the operation finishes.

Complete audio binary, not live playback between nodes

Voice catalog

Get many voices

Retrieve available voice names and IDs as JSON. Use the catalog in a workflow or connect this operation to an AI Agent's Tool input.

Voice catalog as structured JSON

How to set up Voicely in n8n

Install the community node on your self-hosted n8n instance, then connect your Voicely account. Store your API key in n8n credentials.

Package version shown here: 0.1.1. For the WebSocket action, use a current n8n runtime with Node.js 22 or newer and native WebSocket support.

n8n community-node installation instructions

  1. Install the package

    As an n8n owner or administrator, open Settings → Community nodes → Install. Enter this exact package name and follow n8n's installation acknowledgement.

    @voicelystudio/n8n-nodes-voicely

    npm package · Source code and README

  2. Connect your Voicely API credential

    Create a Voicely API key. In the node, create a Voicely API credential, paste the key into its masked field and use this Service URL. Save and test the connection.

    https://n8n.tryvoicely.com

    Fund your API wallet or enable conversion from purchased credits in your API access settings. A successful credential test confirms account access, not funding.

  3. Map your text and generate speech

    Add a manual trigger and Voicely. Select Speech → Generate, enter text or map a previous node's field, pick a voice and model, and set Language Code to match your script. Select MP3 and execute the workflow.

    {{ $json.text }}

    This expression reads a field named text from the incoming item. For a manual test, type a short script directly instead.

  4. Use the audio binary in your next node

    Open Output → Binary → audio. Preview or download speech.mp3, then pass the audio field to a downstream node that accepts binary input. Select Generate via WebSocket to use that transport explicitly.

Import an n8n text-to-speech workflow

Start from the public JSON examples. They contain no credentials or saved execution data. Open the raw JSON, copy it, and paste it onto an empty n8n canvas. Select your credential before executing.

HTTP speech workflow

A manual trigger and a Voicely speech node. Set your credential and execute to create an audio file.

Open http speech workflow JSON

WebSocket speech workflow

The same audio-file workflow with the WebSocket operation selected. Use a current n8n host with native WebSocket support.

Open websocket speech workflow JSON

AI Agent voice discovery

A wired Agent, chat-model node and Voicely voice-catalog tool. Add your Voicely and separate chat-model credentials.

Open ai agent voice discovery JSON

Build voiceover automation around your existing tools

These are workflow recipes you assemble with your own source and destination nodes, not additional prebuilt Voicely connectors.

Spreadsheet scripts to voiceovers

Approved spreadsheet row → Voicely → storage

Map a text column into the speech action. Pass the returned audio binary to a storage node and save the destination URL with the original row.

Articles to audio narration

CMS article → prepared narration → Voicely → upload

Use approved article text as narration, or prepare it with your own text-processing step. Generate an MP3 and upload it using your existing publishing workflow.

Multilingual content automation

Localized script + language code → Voicely → audio file

Provide the translated script and its language code for each item. Keep the original content ID with the returned audio to organize localized voiceovers.

Multilingual text to speech for your n8n workflows

Choose from 31 API languages, derived from the same catalog as Voicely Studio. Set Language Code explicitly to match the supplied text. Preview labels below distinguish preview-only languages and preview regional locales.

  • Arabic

    ar-EG

  • Bangla

    bn-BD

  • Chinese MandarinPreview

    zh-CN (preview), zh-TW (preview)

  • Dutch

    nl-NL

  • English

    en-IN, en-US, en-GB (preview), en-AU (preview)

  • French

    fr-FR, fr-CA (preview)

  • German

    de-DE

  • GujaratiPreview

    gu-IN (preview)

  • HebrewPreview

    he-IL (preview)

  • Hindi

    hi-IN

  • Indonesian

    id-ID

  • Italian

    it-IT

  • Japanese

    ja-JP

  • KannadaPreview

    kn-IN (preview)

  • Korean

    ko-KR

  • MalayalamPreview

    ml-IN (preview)

  • Marathi

    mr-IN

  • Polish

    pl-PL

  • Portuguese

    pt-BR, pt-PT (preview)

  • PunjabiPreview

    pa-IN (preview)

  • Romanian

    ro-RO

  • Russian

    ru-RU

  • Spanish

    es-ES, es-MX (preview), es-419 (preview)

  • SwahiliPreview

    sw-KE (preview)

  • Tamil

    ta-IN

  • Telugu

    te-IN

  • Thai

    th-TH

  • Turkish

    tr-TR

  • Ukrainian

    uk-UA

  • Urdu

    ur-PK

  • Vietnamese

    vi-VN

Use the language and locale reference for API code selection, or explore Voicely's language guides.

Give your n8n AI Agent a voice-catalog tool

Connect Voicely to the Agent's Tool input and choose Voice → Get Many. The Agent receives voice names and IDs as JSON and can use them in its answer. The public example is wired for this action.

Connect your own tool-capable chat model separately. Generate playable audio with the speech action in the main workflow, where the next node receives its binary file.

Open the AI Agent tool example

Choose the right streaming transport

HTTP Generate starts with a complete script and returns a complete file. Generate via WebSocket receives audio frames during synthesis, then returns a complete file as well. Downstream n8n nodes work with the finished binary, not live audio chunks.

Building an application that needs incremental text input and audio-frame playback? Use the WebSocket text-to-speech API guide for that separate application architecture.

Your Voicely account, API credits and spend controls

The node uses the existing Voicely paid API. Fund the API wallet, or enable conversion from purchased credits within the daily limit you set on API access. There is no separate Voicely n8n subscription or second integration charge. Your n8n hosting and connected chat-model costs are separate.

HTTP Generate supports an Idempotency Key for the same job. WebSocket sessions do not have HTTP replay protection; inspect uncertain failures before retrying because completed segments can already be charged. Keep Retry On Fail disabled for uncertain WebSocket sessions.

Manage API funding and spending · Read API billing details

Questions about the Voicely n8n integration

How do I add text to speech to n8n?

Install @voicelystudio/n8n-nodes-voicely from Settings > Community nodes > Install on a self-hosted n8n instance. Add Voicely to a workflow, create a Voicely API credential, choose Speech > Generate, enter or map your text, select a voice and execute. The Binary tab contains the generated audio.

Is the Voicely node available on n8n Cloud?

Published for self-hosted n8n. Submitted to n8n for verification; review is pending. n8n Cloud installation is not available yet.

Which URL and API key should I use?

Create a key on Voicely's API access page. Store it in an n8n Voicely API credential and use https://n8n.tryvoicely.com as the Service URL. The credential connection test checks account access; speech generation also needs paid API funding.

Can n8n generate an MP3 with Voicely?

Yes. Select MP3 as the output format. The node returns speech.mp3 as audio/mpeg in the audio binary field by default. A downstream node can upload, store or send that file. Raw PCM and mu-law formats are also selectable; raw output is not a WAV file.

Does WebSocket stream audio live between n8n nodes?

The WebSocket action receives audio chunks from Voicely, then returns a complete binary file to the next node. Choose Generate for HTTP or Generate via WebSocket explicitly. WebSocket is not a promise of faster workflow completion. It requires a current n8n runtime with Node.js 22 or newer and native WebSocket support.

How do I select a language, voice and model?

Choose a voice and model from the populated dropdowns. Set Language Code to the API language or locale code that matches your script. The catalog on this page marks preview languages and preview locales. Supply your own translated text; selecting a language does not translate it.

Can an n8n AI Agent use Voicely as a tool?

Yes. Connect Voicely to the Agent's Tool input and choose Voice > Get Many to retrieve voice names and IDs as JSON. Connect a separate tool-capable chat model to the Agent. For playable audio files, use the HTTP or WebSocket speech action in the main workflow.

How is the n8n integration billed?

Speech generation uses Voicely's existing paid API billing. Fund your API wallet, or enable conversion from purchased Voicely credits within the daily spending limit you choose on API access. There is no separate Voicely n8n billing plan or second integration charge. Your n8n hosting and any connected chat model are separate.

Can I retry a failed workflow?

Check the returned error before retrying. HTTP Generate supports an Idempotency Key for replaying the same job. WebSocket does not provide HTTP idempotent replay; completed segments can already be charged if a later part fails. Keep Retry On Fail disabled for uncertain WebSocket sessions and inspect the result first.

Put text-to-speech inside your next n8n workflow

Connect your Voicely account, choose a voice and turn your workflow's text into an audio file.

Get a Voicely API keyView the npm package