Voice Cloning Troubleshooting

Fixes for the most common voice cloning problems: the model won't load, the microphone is blocked, the clone sounds noisy or robotic, or generation is slow.

Clone your voice free

The model doesn't load or the download stops

The first download is about 200 MB per language. On fast Wi-Fi that takes under a minute; on slow mobile data it can take several minutes. The Studio is built to cope with bad connections:

  • Automatic retries. If the connection drops or stalls for 30 seconds, the Studio waits a moment and retries, up to four times, continuing from the last byte it received where the server allows it.
  • Nothing is wasted. Each model file is saved as soon as it finishes. If the download finally gives up, press Try again: only the missing files are downloaded.
  • Offline? If your device goes offline, the download resumes by itself when the connection comes back.

If it still won't load, work through these messages and fixes:

You seeWhat it meansFix
“Your connection looks slow” or “Data Saver is on”Your browser reports a slow or metered connectionSwitch to Wi-Fi if you can. The download still works, it just takes longer
“The download stalled” or “was interrupted”The connection kept dropping, even after retriesMove closer to the router or switch networks, then press Try again
“Less than 260 MB of free storage” or “storage is full”The browser can't save the modelFree up space on the device (photos, apps, other sites' data), then try again
“Could not save the model… will download again next visit”Private/incognito window, or storage is restrictedUse a normal window if you want the model kept
“Ran out of memory while starting the voice model”The device doesn't have enough free RAMClose other tabs and apps and try again. Phones with less than 4 GB of RAM may not manage; use a laptop or desktop
“Could not start the voice engine”The browser is too old or blocks WebAssembly workersUpdate the browser, or try the latest Chrome, Edge, Firefox or Safari
“Download failed (HTTP 4xx/5xx)”The model server is unavailableWait a few minutes and try again. If it persists, tell us

Still stuck? Click Delete downloaded models in the Studio sidebar, reload the page and try again. This clears any half-saved files.

Older laptops and low-end phones

  • Loading takes a while after the download. “Starting the model…” prepares the AI in memory. On older devices this can take 10–30 seconds; let it finish.
  • Generation is slower than playback. On slow processors you may hear short gaps while audio is generated. Generate shorter texts, or wait for the whole result and play the downloaded file.
  • The tab reloads by itself. On phones this usually means the system ran out of memory. Close other apps, keep only the Studio tab open, and generate shorter texts.

“Microphone access was blocked”

Your browser asked for permission and it was denied. Click the camera/microphone icon in the address bar, allow the microphone and try again. On macOS also check System Settings → Privacy & Security → Microphone. Alternatively, record with another app and upload the file.

“That clip is too short”

After silence is trimmed, at least a few seconds of speech must remain. Record 10–20 seconds of continuous speech.

The clone sounds noisy, echoey or muffled

The model reproduces your reference clip exactly, including its problems. Record again following the recording guide: a quiet, soft room and a steady microphone distance.

The clone doesn't sound like the speaker

  • Use a longer clip (10–20 seconds) with only one speaker.
  • Avoid clips with music, effects or phone-call quality.
  • Make sure the language selected in the Studio matches the language you're generating.

Words are mispronounced

Spell out numbers, abbreviations and unusual names phonetically ("Nguyen" → "Win"). Add a comma before a difficult word to slow down the delivery.

Generation is slow

  • Use the dedicated Studio page: it can use several processor cores and is usually 2–4× faster than the home page demo.
  • Close heavy tabs and apps while generating.
  • On phones, generate shorter texts.

My saved voices disappeared

Voices are stored in your browser on this device. Clearing site data, using a different browser or a private window means they aren't available. Keep your original recordings so you can re-create voices any time.

Frequently asked questions

Why can't I select some languages?

Guests can use English. Create a free account to unlock Spanish, German, Italian and Portuguese.

Ad space · inline