n8n text to speech with a self-hosted voice
n8n is a popular way to automate content: a new post, a row in a sheet or an AI summary goes in, and something useful comes out. Speech is a natural next step: a narration for every article, a voice-over for every short, a daily briefing as audio.
The free SuperClonica engine runs as a small HTTP server next to n8n and turns text into natural speech on your own hardware. This guide shows how to connect the two with the built-in HTTP Request node and what to do with the audio.
What you can automate
n8n moves the text; SuperClonica turns it into speech on your own server, with no per-character bill.
- A narration for every new blog post or RSS item.
- Voice-overs for short videos from a script column in a spreadsheet.
- A daily briefing or summary written by an AI model, read aloud.
- Announcements and notifications as audio.
Choose the voice free
On the page of any voice, paste a typical sentence from your workflow, 30 to 100 characters, choose the language and press Voice it. In a few seconds you hear how your automation will sound, with no account.
Connect n8n to the speech server
The node then hands the next node a 16-bit mono WAV file.
- Start the server: docker run -d --name tts -p 8000:8000 -v superclonica-models:/models superclonica/superclonica:cpu. If n8n runs in Docker too, put both containers on one Docker network and use the container name, http://tts:8000, instead of localhost.
- Add an HTTP Request node: method POST, URL http://tts:8000/synthesize, Send Body on, body content type JSON.
- In the body, take text from the previous node, set voice to F1 or to the content of a voice file, and lang to en.
- Under Options, set Response Format to File and enter the field name for the output, for example audio.
Store or publish the audio
Save the file with the Read/Write Files from Disk node, upload it to cloud storage, attach it to a message, or pass it on to the step that builds your video. For MP3, a self-hosted n8n can run ffmpeg in an Execute Command node. Send long texts paragraph by paragraph or raise SUPERCLONICA_MAX_TEXT, 5000 characters by default. If the server is reachable from other machines, set SUPERCLONICA_API_KEY and send it in the Authorization header as a Bearer token.
Voices and honesty
The ten stock voices of the model, M1 to M5 and F1 to F5, are free; for a voice of your own, train one in the SuperClonica studio from a recording of a speaker who agreed, or order it on the exchange. Say that the audio is machine-generated where that matters to listeners, and never automate speech in a voice that imitates a real person.