How to Download a Microsoft Sam Voice Clip as WAV
Download SAM-style speech as WAV, check its audio format, export separate dialogue lines, and resolve common browser and editor problems.
In this guide
Download a generated voice clip
Open the WAV generator, type your text, click Generate, listen and choose Download WAV. The result is a recording of that line. It is not an installer for the original Microsoft Sam voice.
The site uses Software Automatic Mouth through sam-js. A Microsoft Sam-style label describes the intended retro character; it does not mean the original Windows voice engine is running in your browser.
Make and check one file first
Start with System ready. Proceed to platform three. Keep Speed 72, Pitch 64, Mouth 128 and Throat 128 for this first take. After generating, listen for the complete last word and check the spelling of any names.
Click Download WAV and open the newest file from your browser's Downloads list. If you changed the text or settings after previewing, generate again first: the download uses your current input, which can differ from the last audio you heard.
System ready. Proceed to platform three.Speed 72 · Pitch 64 · Mouth 128 · Throat 128
Reference take. Compare every change against this same sentence.
Load this exampleKnow what is inside the WAV
The current exporter writes a single audio channel at 22,050 samples per second, using 8-bit PCM. WAV is the container; these values describe the audio inside it. No microphone recording, music or sound effect is included.
If your editor needs 48 kHz or a different bit depth, open the file in an audio editor and convert a copy. Keep the original for reference. Conversion can meet an import requirement but does not add missing detail or make the voice sound like modern neural TTS.
Scroll the table sideways on smaller screens.
| Property | Current export |
|---|---|
| Container | WAV |
| Encoding | PCM |
| Channels | 1 (mono) |
| Sample rate | 22,050 Hz |
| Bit depth | 8-bit |
Export dialogue without losing track of takes
For a small set of lines, put each one on a new line and use Batch line rendering. The tool accepts up to twelve lines per batch within the text box's 900-character limit. Allow multiple downloads if the browser asks; otherwise save the lines individually.
Name each result by its job: scene-02-terminal-open-s72-p64.wav is easier to replace than download-7.wav. Record Mouth and Throat too when they differ from 128. Copy Link saves the script and settings in a URL; it is a recipe for regeneration, not an uploaded audio file.
Resolve common download problems
Scroll the table sideways on smaller screens.
| Symptom | Next action |
|---|---|
| The browser saved nothing | Check download permissions and the Downloads list. Test one short clip before retrying a batch. |
| A name sounds wrong | Test the name alone, rewrite it phonetically, then put it back in the sentence. |
| The editor refuses the file | Open it in an audio editor and export the PCM format your destination requires. |
| The wrong take plays | Check the newest filename and regenerate after changing the current input. |
Keep an editable source
Trim only the unwanted silence and leave the final consonant intact. Put pauses between dialogue turns on the editing timeline, where their duration can be controlled exactly. Apply radio noise, reverb and music to a copy so the dry voice remains available.
Before exporting the final video or game build, listen to the line inside the full mix. A file that is clear alone may be masked by music. Lower the competing sound around the instruction rather than relying only on a louder voice.