Agent speaking turns one photo of you and a short script into a video of you talking to camera. It speaks in your own recorded voice or a preset voice. Use it for quick "just listed", open house or market update videos when you don't have time to film.
Before you start
You need:
- A clear, front-facing photo of yourself with your head and shoulders in frame. An avatar you made in Create avatar works well.
- A script of 12 to 70 words.
- A voice. Record your own speaking voice once (it takes about 30 seconds and is free), or use a preset: Charlie, Noah, Michael or Ethan.
It works in two steps, and each is paid separately. First Agentr.ee reads your script out loud so you can hear it. Then it makes the video of the take you choose. Both are priced on the measured length of the finished audio, which runs from 4 to 30 seconds:
| Free | Pro | |
|---|---|---|
| Audio (step one) | 2 credits a second | 2 credits a second |
| Video (step two) | 35 credits a second | 35 credits a second |
For example, a 12-second take costs 24 credits for the audio and 420 credits for the video, 444 credits in all. The shortest possible video comes to about 148 credits, and a full 30 seconds to about 1,110. Pro agents pay credits for these videos too, from their monthly allowance.
How to create a talking video
- Open AI Studio at agentr.ee/workspace/studio and choose Agent speaking under Create videos.
- Under 1 · Your photo, press Choose a photo and pick one from your Media Library.
- Tick Make me walk if you want to walk toward the camera while you talk. This changes the video only.
- Under 2 · Your script, write what you want to say, or tap a starter such as Just Listed or Open House and fill in the blanks yourself.
- Under 3 · Voice, choose Your speaking voice, or Record your speaking voice to set it up, or pick a preset.
- Under 4 · Shape, choose Vertical · 9:16, Square · 1:1 or Landscape · 16:9.
- Press Generate audio. It takes a few seconds.
- Play the take under Your takes. Press Read it again if you want another reading.
- Choose Use this on the take you like. The credits to film it show beside it.
- Press Make the video. There is no second confirm step.
The video takes a few minutes. You can close the page and come back.
What you get
You get an MP4 of you speaking your script: 720×1280, 960×960 or 1280×720, depending on the shape. Press Download on the result screen, or Make another. You can also find it later on the Generations page under the Video filter. The audio takes can be played but not downloaded.
Tips for better results
- Hear the audio before you film. If the wording sounds off, change the script before you pay for the video.
- You can keep up to 5 takes. A new reading after that replaces your oldest take, never the one you chose.
- Keep scripts short. Seventy words is about as long as anyone watches.
- Check the estimate under your script. The final price is measured from the finished audio.
- Tick or untick Make me walk as often as you like. It never costs a new reading.
Common questions
Does it cost more if I generate the same audio twice?
No. If you press Generate audio again with the same script, voice and photo, you get the take you already have at no charge. Read it again always makes a new reading and charges for it.
What happens to my credits if it fails?
If the audio or the video doesn't come back, the credits for that step come back automatically. Your takes are kept, so you can try the video again without paying for the audio again.
Is Agent speaking free on Pro?
No. It costs the same credits on every plan.
How do I record my own voice?
Pick Record your speaking voice under 3 · Voice and read the short script out loud. It needs 10 to 30 seconds and saves as soon as you stop.
Related articles
Read How to teach AI your writing and speaking voice, How to create an AI headshot with Create avatar and How AI credits work in AI Studio.