Speaking practice with a face
Practise the conversation before it counts.
Talk face to face with a live AI partner: a job interview, a salary talk, a sales call. Then read a scored report on how you sounded. Free, no sign-up.



Mira, Sonia or Sam will play the other side, live on video.
Pick one to begin
Job interviewA hiring manager asks real questions and follows up when you are vague. Salary talkA recruiter makes an offer and pushes back when you ask for more. Sales callA busy buyer takes your call, raises objections and decides on a next step. Hard conversationGive tough feedback, ask for a raise, or say no. They react like a person would. Upset customerA customer calls with a real problem and a short fuse. Calm them down and fix it. Speaking practiceAn easy conversation on a topic you pick, for fluency and confidence. Your ownDescribe any conversation and who you need them to be.How it works
- 1
Say what you are preparing for
Pick the conversation and add a few details: the role, the offer, what you sell. Your partner plays the other side.
- 2
Talk for 3 minutes
They appear on video, ask real questions and react to what you say. You answer out loud, like the real thing.
- 3
Read your report
A score, five measured numbers, what worked, what to fix, and one of your answers rewritten tighter.
What the report measures
The numbers come from the timing of every word you said. They are counted, not guessed.
How the scoring works- Pace
Words a minute while you were talking. Too fast and points blur, too slow and people drift.
- Filler words
Every um, you know and sort of, counted and marked in the transcript.
- Talk balance
Your share of the talking against what suits that kind of conversation.
- Answer length
Your longest answer and the average, so you can see where you rambled.
- Response time
How long you took to start answering, and how often you cut in.
- Written feedback
Two or three things that worked and two or three to change, quoting your own words, plus a tighter version of your weakest answer.
Questions
Is it free?
Yes. You get 1 rehearsal a day, 3 minutes, with no sign-up.
Am I talking to a real person?
No. Your partner is an AI with a face and a voice, playing a part you choose. It says so if you ask.
What do I need?
A microphone and a recent browser. A camera is optional and off unless you switch it on.
What happens to what I say?
The words are kept as a transcript so your report can be built, and deleted after a set time or when you press Delete. We do not keep audio or video. The privacy page has the details.
Why is the room sometimes busy or closed?
There is one room, and one person can use it at a time. If someone is in it, you see how long until it frees up. The room also has a small amount of video time each day, shared by everyone. When that is used up, it opens again at midnight UTC.
Which languages work?
English only for now, both for what you say and what your partner says.