Chatter Box: Voice Cloning with TTS

Upload a short voice clip and input text to generate new speech in the cloned voice.

Download Workflow Try in Cloud

Input

Text to speak

Voice sample

0:00 / 0:00

Exaggeration

Default: 0.5

Advanced settings

CFG weight

Default: 0.5

Seed

Temperature

Default: 0.8

Run

Output

Example

00:00

Call this model from your code

The request below uses the same prepared graph and settings as the Playground. Each submission can start a paid run; do not automatically retry a submission whose outcome is unknown.

Get API key

What you need

Runs on: Comfy Cloud
Key: COMFY_API_KEY
Your files: Uploaded by the code from their urls

Needs a paid Cloud plan and credits.

Download API JSON

# Python 3.10+: pip install comfy-sdk==0.4.0
import os

from comfy_sdk import Comfy

api_key = os.environ.get("COMFY_API_KEY", "").strip()
if not api_key: raise RuntimeError("Set COMFY_API_KEY")
os.environ.setdefault("COMFY_BASE_URL", "https://cloud.comfy.org")
client = Comfy(api_key=api_key)
workflow = client.workflows.from_json({
    "4": {
        "class_type": "FL_ChatterboxTTS",
        "inputs": {
            "audio_prompt": [
                "6",
                0
            ],
            "cfg_weight": 0.5,
            "exaggeration": 0.5,
            "keep_model_loaded": False,
            "seed": 390519100,
            "temperature": 0.8,
            "text": "Good evening, and welcome aboard the late train to the coast. Our journey tonight will take a little under two hours, with stops at Millbrook, Harrow Bridge and the old harbour town of Kestle. The buffet car is open until ten, serving hot tea, soup and fresh bread from the bakery at the station. If you look out of the left windows as we cross the river, you may see the herons fishing in the shallows. Please keep your tickets close at hand, settle in, and enjoy the ride.",
            "use_cpu": False
        },
        "_meta": {
            "title": "FL Chatterbox TTS"
        }
    },
    "5": {
        "class_type": "SaveAudioMP3",
        "inputs": {
            "audio": [
                "4",
                0
            ],
            "filename_prefix": "audio/chatterbox_tts",
            "quality": "V0"
        },
        "_meta": {
            "title": "Save Audio (MP3) (DEPRECATED)"
        }
    },
    "6": {
        "class_type": "LoadAudio",
        "inputs": {
            "audio": "UPLOADED_voice_FILENAME"
        },
        "_meta": {
            "title": "Load Audio"
        }
    }
})
workflow.set_input("6", "audio", client.assets.from_url("https://comfy.org/catalog-media/a8cb87b0-2285-4cd9-bee8-c0a5011ced19/3bb18f46-c38e-4407-94e0-2a808affc011"))

# run() also passes api_key to the partner nodes in this workflow.
job = client.run(workflow, api_key=api_key)
for output in job.get_outputs("5"):
    output.to_file(output.name)

API documentation

Try an example

About this workflow

The 'Chatter Box: Voice Cloning with TTS' workflow is designed to clone a voice from a short audio clip and generate new speech using text-to-speech (TTS) technology. This workflow utilizes the Chatter Box model, which is specifically tailored for voice cloning tasks. By leveraging the FL_ChatterboxTTS node, users can input a short voice recording, typically between 5 to 10 seconds, and a text prompt to produce speech in the same voice as the input clip. The process is streamlined and efficient, making it accessible for both technical and non-technical users.

Technically, the workflow begins with the **LoadAudio** node, which ingests the user's audio file. This is followed by the **FL_ChatterboxTTS** node, which processes the audio to identify and replicate the unique characteristics of the voice. The text input is then converted into speech that mimics the original voice. The final output is saved using the **SaveAudioMP3** node, allowing for easy access and sharing.

The workflow also includes a **MarkdownNote** node for documentation purposes, ensuring that users can keep track of their inputs and outputs. This workflow is particularly useful for content creators, educators, and developers looking to personalize audio content with minimal effort.

Chatter Box: Voice Cloning with TTS

Input

Inside the workflow

Call this model from your code

What you need

Try an example

About this workflow

Questions & answers

Input

Inside the workflow

Call this model from your code

What you need

Try an example

About this workflow

Questions & answers

What is the minimum length of the voice clip needed for cloning?

Can I use this workflow for commercial projects?

What file format is the generated audio saved in?

Does this workflow support multiple languages?