Migrating from ElevenLabs to LMNT

Congratulations on migrating to LMNT. We’re glad to have you! As you know, LMNT’s TTS model is the best in the industry, with consistently low latency (~150ms between input and output), reliable service, and superb speech quality (we infuse the human “LMNT” in our speech models!). Here’s a guide to help you make the migration.

​

1. Subscribe to LMNT

Visit Pricing on LMNT.com and select your desired plan.

As a thank you for migrating over, we’d be glad to offer you 500,000 free characters for getting started. Please email support@lmnt.com with the email title “Migrated from ElevenLabs to LMNT” and we’ll get it taken care of.

​

2. Download our SDK

We currently support Python, Node, and Unity. Here’s how to use each.

• Python: pip install -U lmnt
• Node: npm install lmnt-node
• Unity: follow the instructions here

​

3. Migrate your code

The code samples below show a before and after of one developer’s code when migrating from ElevenLabs to LMNT. You can also use this GPT we built to help you migrate over your code!

​

ElevenLabs Code

import os
import uuid
from elevenlabs import VoiceSettings
from elevenlabs.client import ElevenLabs

ELEVENLABS_API_KEY = "11 LABS API KEY"
client = ElevenLabs(
   api_key=ELEVENLABS_API_KEY,
)

def text_to_speech_file(text: str) -> str:
   response = client.text_to_speech.convert(
       voice_id="pNInz6obpgDQGcFmaJgB",
       optimize_streaming_latency="0",
       output_format="mp3_22050_32",
       text=text,
       model_id="eleven_turbo_v2",
       voice_settings=VoiceSettings(
           stability=0.0,
           similarity_boost=1.0,
           style=0.0,
           use_speaker_boost=True,
       ),
   )

   save_file_path = "/Users/name/Downloads/11LABS.mp3"

   with open(save_file_path, mode = "wb") as f:
       for chunk in response:
           if chunk:
               f.write(chunk)

   print(f"{save_file_path}: A new audio file was saved successfully!")

   return save_file_path

text_to_speech_file("Hello world.")

​

LMNT Code

LMNT_API_KEY="LMNT API KEY"

import asyncio
from lmnt.api import Speech

text = 'Hello world!'
voice_id = 'ava'
speed = 1.0
language = 'en'

async def main():
 async with Speech(LMNT_API_KEY) as speech:
   synthesis = await speech.synthesize(text, voice_id, speed, language)
   with open('/Users/yourname/Downloads/audio.mp3', 'wb') as f:
     f.write(synthesis['audio'])

asyncio.run(main())

​

4. Select your desired text, voice, speed, and language

You may choose from a stock voice (we love ava and zeke!) or create your own voice clones in our Playground. We have heard that Jennifer → Savannah has been a good switch, as has GuyNeural → Daniel. (And, the Python code below can list all of your available voices.)

LMNT_API_KEY="LMNT API KEY"
import asyncio
import pprint
from lmnt.api import Speech

async def main():
   async with Speech(LMNT_API_KEY) as speech:
       voices = await speech.list_voices()
       name_id_list = [(voice['name'], voice['id']) for voice in voices]

   for name, id_ in name_id_list:
       print(f"Name: {name} | ID: {id_}")

await(main())

If there are any specific voices you’d like to transfer over, contact us at support@lmnt.com and we’ll help you make the transition. If you want to improve any aspect of your application, check out our Documentation for guides for optimizing quality, optimizing latency, overriding pronunciations, and more. And if there are any roadblocks in your way, we’re more than happy to offer more individualized support. You can email us at support@lmnt.com with any questions you may have or book time with someone on our team here.

You’re all set! You should be good to go with your new LMNT-powered application. Welcome to the community!