#i want to use voice changer endpoint but i am getting erros with my cloned voice.

1 messages · Page 1 of 1 (latest)

finite crescentBOT
#
AI-generated Answer for adriguerra27

Thanks for reaching out. Here are the most common causes of Voice Changer (speech‑to‑speech) errors with a cloned voice and how to fix them, based only on ElevenLabs’ official docs.

What to check first

  • Endpoint and required fields: Voice Changer is POST /v1/speech-to-speech/:voice_id with multipart/form-data including the source audio file. The model must support voice conversion (check can_do_voice_conversion), and voice_settings must be JSON-encoded if provided.
  • Model choice: If you see “The selected model can not be used for voice conversion,” reselect a speech‑to‑speech model (e.g., eleven_multilingual_sts_v2 or eleven_english_sts_v2). If switching between TTS and Voice Changer caused it, toggle models and retry.
  • Voice ID and API key: 400/401 errors commonly stem from invalid API key (xi-api-key), wrong voice_id, or quota_exceeded. Verify your key, confirm the voice_id from My Voices or via List Voices, and ensure you have credits.
  • Professional Voice Clone readiness: If you get “No model found for this voice. Please select another voice,” your PVC likely hasn’t finished fine‑tuning. Wait until its status shows ready (typically 3–6 hours, up to 24), and check status per‑model in My Voices.
  • Request formatting (422): A 422 means a required value is missing/invalid. In JS, 422 “cors” often occurs if you don’t JSON.stringify() the body when required. For Voice Changer specifically, send multipart/form-data, ensure -F audio=@file is set, and JSON‑encode voice_settings if used.
  • Audio constraints: Input length is max 5 minutes (300s). Use a supported format (audio: MP3, M4A, FLAC, OGA/OGG, WAV; video: MKV, WEBM, MP4, MOV). Exceeding limits or unsupported formats can cause failures.
  • Credits: Voice Changer costs 1,000 credits per minute; insufficient credits can return quota_exceeded.

Known‑good request template (from docs)

  • POST https://api.elevenlabs.io/v1/speech-to-speech/{voice_id}
  • Headers: xi-api-key: YOUR_KEY, Content-Type: multipart/form-data
  • Form fields: audio=@yourfile, model_id=eleven_multilingual_sts_v2 (or eleven_english_sts_v2), optional voice_settings as a JSON string, optional remove_background_noise=true, optional file_format=pcm_s16le_16 for raw 16‑bit 16kHz mono PCM.

If you share:

  • The exact error message and HTTP status
  • The model_id you’re using
  • Whether the voice is an Instant Voice Clone or a Professional Voice Clone and its readiness status
  • A redacted snippet of your request (headers and form fields)

…I can map it to the precise fix per the docs above.