Start a separate project
Sign in, open Projects and choose New project. Give it a name you can recognise, such as Speaking practice. This keeps your first exercise separate from other work. As a guest, you can explore the demo before creating an account.
Script: check a small set of lines
Write a few lines that you want in the recording. These original examples give you a short production exercise:
I prepare the point I want to make.
I give myself time to finish my thought.
I speak at a pace that feels comfortable.
Put each sentence on its own line. Read the text aloud and remove headings or notes from the spoken script. If you use AI suggestions, review their meaning before continuing. The example is a writing exercise, with no promised personal outcome.
Voices: listen before mixing
Choose a voice that supports your script language. A preview lets you check the delivery; it does not read your own text. Check the generation estimate before starting paid speech generation. Split shares the script across the cast; Full gives each voice a complete reading and can cost more. Once the voices are ready, play a few of your own lines on their own from the Voices dock and listen for mispronounced names or words before you add music.
To record instead, add Your Voice and follow the recording prompts. Listen for missed words or distortion. The own-voice guideexplains how to prepare a take.
Music: choose a bed you want to hear
Add a library bed or an audio file you have permission to use. Play several sections and check for abrupt level changes. Keep the first mix simple so you can identify which layer needs attention.
Mix: compare the separate parts
Listen to the voice alone, then with the bed. Use the level controls to place the voice beneath the background while keeping your listening volume comfortable. Check a quiet passage of the bed as well as a louder one. The masking guide explains the suggested range.
Optional layers can wait until you understand the basic mix. Leave them out of your first exercise if you do not yet know what they add.
Export: save and check the result
Check the file name and track length. Choose Studio master (WAV) to preserve all rendered layers, or Compact copy (MP3) for a smaller ordinary listening file. If you use the silent layer, follow the format guide.
Use the Render button for the chosen format. Regular rendering does not use credits; rendering on the studio server is a separate paid option. If the button is locked, complete the step named beside it.
Open the finished download in a player. Check the beginning and ending, confirm the duration and try repeat playback if you intend to loop it. Keep the script and editable project so you can revise the recording later.