Fluent Me
Practice

YOUR ENGLISH PRACTICE

Make your next conversation easier.

A little practice today. More confidence out there.

Sign in to keep your progress
YOUR AI CONVERSATION COACH5 min

Your words.
A little more confident.

A real conversation, useful feedback, and a chance to try again.

Video coach
Speak or type

Sign in with Google to start practising

Practise for real life

Choose a starting point

Recent practice

Your practice recaps will appear here after your first conversation.

SET UP YOUR PRACTICE

What would you like to practise?

Choose a starting point. Your microphone stays off until you’re ready.

Conversation mode
Session length

Timed sessions give you a one-minute warning, then close with a recap. You can still finish early.

WeChat audio bridge

Choose the virtual input carrying the WeChat caller into the twin, and the virtual output carrying the twin back to WeChat.

Virtual devices appear after their driver is installed. Browser permission may be needed before names are visible.
Your coach is readyNo room time starts until you continue.

OPTIONAL PERSONALIZATION

Create your coach

Practice face to face with a coach that uses your own face and voice—or keep the stock male coach and start immediately.

Stock coach active
Your recording begins only when you press Record.

Video and audio stay on this device until you explicitly create a voice clone or submit a hosted training URL. Nothing is uploaded merely by opening this setup.

What would you like to personalize?
IDENTITY & CONSENT

Only create a coach from your own likeness and voice, or content you have explicit permission to use.

Read this at the beginning of the recording Your name will replace the bracketed placeholder before recording.
CAPTURE ON THIS DEVICE

Record one clean minute.

This browser capture is a convenient draft/export. Tavus recommends a desktop recorder or PAL Maker for reliable 1080p at 25 fps or higher.

Camera is off Press Record when you are framed and ready.
00:00 / 01:00
  1. 0–8 secRead the consent phrase clearly
  2. 8–30 secSpeak naturally with minimal movement
  3. 30–60 secStay still, lips closed, eyes on camera
No recording yet
YOUR VOICE · ELEVENLABS

Clone the voice.

Record a separate voice sample so the silent half of your Tavus face video never dilutes voice quality.

VOICE SAMPLE Speak continuously for 60–90 seconds. Use a quiet room, stay close to the microphone, and vary your pace naturally. This recording stays on this device until you press Create my voice clone.
Plan
Checking…
Character usage
Checking…
Voice limit
Checking…

Nothing uploads before this step. Pressing it explicitly sends the reviewed sample to ElevenLabs. Connecting that private voice to your video coach also gives Tavus the dedicated ElevenLabs integration key required for synthesis.

Record and review a sample to continue.
YOUR FACE · TAVUS

Connect your trained face.

Use Tavus PAL Maker for guided recording and quality checks. Training usually takes 3–4 hours; return here with the Face ID when it is ready.

Recommended: train in PAL MakerRecord, pass the quality checks, and let Tavus train your Face.
Open PAL Maker
Advanced: submit a hosted training video

The Tavus API needs a direct HTTPS download that remains valid for at least 24 hours.

Train in PAL Maker, then connect the Face ID here.
Use either model—or combine both Your voice can use the stock male face; your Face can use the stock voice. The stock coach always remains available.

Using the stock coach clears this browser's saved selection. It does not delete your Tavus Face, ElevenLabs voice, or personal PAL from either provider.

YOUR SPACE TO GROW

Keep your practice
with you.

Sign in to practise with your coach and bring your progress to every device.

By continuing, you agree to the beta terms. See how your practice is handled in our privacy notice.