danny-avila/LibreChat · warning

Voice ${voice} is not available.

Error message

Voice ${voice} is not available.

What it means

openAIProvider rejects a voice that is not in ttsSchema.voices, unless the list contains the sentinel 'ALL'. The check only runs when `voices` is a non-empty array, so an unset/empty voices list disables the allowlist (any voice passes). A configured list thus acts as a strict gate.

Source

Thrown at api/server/services/Files/Audio/TTSService.js:119

  /**
   * Prepares the request for OpenAI TTS provider.
   * @param {Object} ttsSchema - The TTS schema for OpenAI.
   * @param {string} input - The input text.
   * @param {string} voice - The selected voice.
   * @returns {Array} An array containing the URL, data, and headers for the request.
   * @throws {Error} If the selected voice is not available.
   */
  openAIProvider(ttsSchema, input, voice) {
    const url = ttsSchema?.url || 'https://api.openai.com/v1/audio/speech';

    if (
      ttsSchema?.voices &&
      ttsSchema.voices.length > 0 &&
      !ttsSchema.voices.includes(voice) &&
      !ttsSchema.voices.includes('ALL')
    ) {
      throw new Error(`Voice ${voice} is not available.`);
    }

    const data = {
      input,
      model: ttsSchema?.model,
      voice: ttsSchema?.voices && ttsSchema.voices.length > 0 ? voice : undefined,
      backend: ttsSchema?.backend,
    };

    const apiKey = resolveConfigSecret(ttsSchema?.apiKey) || '';
    const headers = {
      'Content-Type': 'application/json',
      ...(apiKey && { Authorization: `Bearer ${apiKey}` }),
    };

    return [url, data, headers];
  }

View on GitHub (pinned to 5ff282f900)

Solutions

1. Use a voice id returned by GET /api/voices for the current provider.
2. Add the desired voice id to ttsSchema.openai.voices in librechat.yaml.
3. If you want to allow any voice, include 'ALL' in the voices list (or leave voices unset).

Example fix

# before
speech:
  tts:
    openai:
      voices: ['alloy', 'echo']
# request uses voice='nova' -> error

# after
speech:
  tts:
    openai:
      voices: ['alloy', 'echo', 'nova']

const allowed = ttsSchema?.voices;
if (Array.isArray(allowed) && allowed.length > 0 && !allowed.includes(voice) && !allowed.includes('ALL')) {
  return res.status(400).json({ error: `Voice ${voice} is not in the allowed list.` });
}

When it happens

Trigger: POST /api/speech/tts with the openai provider and a voice id not present in ttsSchema.openai.voices (and the list does not include 'ALL').

Common situations: Frontend cached an older voice list; config voices changed; voice id typo; user-supplied voice not in the allowlist.

Related errors

• Text is required
• Missing DALLE_API_KEY environment variable.
• Missing required field: prompt
• Missing required field: prompt
• Missing required field: finetune_id for finetuned generation

AI-assisted analysis of danny-avila/LibreChat@5ff282f900 (2026-08-12). Data as JSON: /api/errors/39513a627957f240. Report an issue: GitHub.