Voice Changer

Knox exposes ElevenLabs-compatible Voice Changer (speech-to-speech) endpoints at https://api.knoxstudio.ai/v1 for KnoxStudio. Authenticate with your Knox API key (Authorization: Bearer sk-... or xi-api-key). Knox bills the request and relays it upstream. The same routes are also served by https://api.knox.chat/v1.

This is not Text to Speech, not Realtime speech-to-speech with gpt-realtime-2.1, and not Voice Isolator. Voice Changer takes source audio and restyles it with a target voice_id while keeping emotion, timing, and delivery.

Typical flow:

1. List models — GET /v1/speech-to-speech/models
2. Pick a voice_id from Voices (GET /v1/voices or GET /v2/voices)
3. Convert — POST /v1/speech-to-speech/{voice_id}

https://api.knoxstudio.ai/v1
Authorization: Bearer sk-...

A commonly used public voice is 21m00Tcm4TlvDq8ikWAM (Rachel).

Models and billing

curl https://api.knoxstudio.ai/v1/speech-to-speech/models

model_id	Use	Price
eleven_multilingual_sts_v2	Default multilingual voice changer	$0.17 / audio minute
eleven_english_sts_v2	English-only voice changer	$0.17 / audio minute

Uploads may be up to 500MB. Audio must be at least 100ms and at most 10 minutes. Default model is eleven_multilingual_sts_v2.

Unauthenticated catalog requests are limited to 60/minute and may be cached for 5 minutes. A Bearer token returns only the models that token is allowed to use.

Convert speech to speech

Send multipart/form-data with the source audio. Do not send a TTS model_id such as eleven_multilingual_v2.

Streaming variants do not support WAV output_format values.

cURL

curl -X POST "https://api.knoxstudio.ai/v1/speech-to-speech/21m00Tcm4TlvDq8ikWAM?output_format=mp3_44100_128" \
     -H "Authorization: Bearer $KNOXCHAT_API_KEY" \
     -F "audio=@speech.wav" \
     -F "model_id=eleven_multilingual_sts_v2" \
     -F "remove_background_noise=true" \
     -o changed.mp3

Python

import os
import requests

voice_id = "21m00Tcm4TlvDq8ikWAM"
with open("speech.wav", "rb") as audio:
    response = requests.post(
        f"https://api.knoxstudio.ai/v1/speech-to-speech/{voice_id}",
        headers={"Authorization": f"Bearer {os.environ['KNOXCHAT_API_KEY']}"},
        params={"output_format": "mp3_44100_128"},
        files={"audio": audio},
        data={
            "model_id": "eleven_multilingual_sts_v2",
            "remove_background_noise": "true",
            "voice_settings": '{"stability":0.5,"similarity_boost":0.75}',
        },
    )
response.raise_for_status()
open("changed.mp3", "wb").write(response.content)

Useful fields: file_format (pcm_s16le_16 or other), seed (0–4294967295), voice_settings (JSON object), remove_background_noise (true / false). Optional query: enable_logging, optimize_streaming_latency (0–4, deprecated).

Related endpoints

Goal	Endpoint
Stream converted audio	POST /v1/speech-to-speech/{voice_id}/stream
List / search voices	GET /v1/voices, GET /v2/voices

Errors

422 uses { "detail": [{ "loc": [...], "msg": "...", "type": "value_error" }] }. Other errors use { "detail": "..." }.

Common statuses: 401 missing/invalid key, 402 insufficient credits, 403 model not allowed, 422 missing audio, over 10 minutes, invalid STS model_id, or invalid output_format.