Zum Hauptinhalt springen

AI Audio Generation

Instead of sourcing and hosting your own audio files, you can generate text-to-speech announcements and background music for voice support directly from the Voice Support dashboard. Generated clips are stored in your server's file library, show up in the audio picker for your waiting and closed-state playlists, and are charged in AI Coins from your server's balance.

Where to find it​

Any time the dashboard shows an audio picker - for the waiting music or closed-state music playlist - you'll see a Generate (AI) tab next to Upload and Library. That tab splits further into TTS (text-to-speech) and Music.

Both tools work the same way:

  1. Fill in your prompt (text for TTS; a description for music).
  2. Review the live cost estimate in AI Coins and your current balance.
  3. Click Generate. The wait is typically a few seconds for TTS, up to a minute for music.
  4. Preview the result with the inline play button.
  5. The clip is automatically registered in your file library and added to the playlist you're editing. If you prefer, you can also upload the clip and add it manually later.
Generated clips don't count against your storage quota

Because AI-generated tracks are paid for with AI Coins, they don't take up any of your server's file-storage quota - no matter how many you generate, you won't run out of storage space from AI audio. Tracks you upload yourself do count against the quota like any other file.

Text-to-speech (TTS)​

Useful for greeting messages, queue-closed announcements, hold music voiceovers, and any other spoken content. Produced with ElevenLabs voices.

How it works​

  1. Pick a language. The picker defaults to your browser language if available, and shows a count next to each language.
  2. Optionally enable Include multilingual voices to add voices that work across any language (handy for mixed-language queues).
  3. Pick a voice. Each voice card shows the voice name, an accent label, a short use-case tag, and a preview button. You can listen to the preview before selecting it.
  4. Type the text you want spoken - up to 10,000 characters.
  5. Click Generate. The cost is calculated from the text length (see pricing below).

Pricing​

ItemValue
Cost rate1 AI Coin per 4 characters (rounded up), with a minimum of 1 coin.
Maximum text length10,000 characters per generation.

The exact cost is shown live as you type, and the button is disabled if your balance is insufficient. A "Buy more coins" link appears when you're short.

Music​

Useful for producing waiting music that matches your server's vibe without sourcing licensed tracks. Produced with ElevenLabs' music model.

How it works​

  1. Write a prompt describing the vibe, instrumentation, tempo, and mood - e.g. "warm lo-fi beat, mellow piano, soft drums, 70 BPM, no vocals". Up to 4,100 characters.
  2. Pick a duration between 3 seconds and 5 minutes (default: 60 seconds).
  3. Click Generate. Music generation takes longer than TTS - the UI shows a progress indicator.

Pricing​

ItemValue
Cost rate12 AI Coins per second (= 600 coins per minute).
Duration range3 seconds to 300 seconds (5 minutes) per generation.

Again, the live cost preview updates as you drag the duration slider, and the button is disabled if you can't afford the generation.

Tips​

  • Build a library over time. Generations are saved to your file library even after you close the dashboard. You don't need to re-generate a greeting every time - reuse the stored clip.
  • Mix AI and uploaded tracks. The playlist is just a list of URLs. You can freely mix AI-generated tracks with files you uploaded yourself.
  • Preview before adding. The inline preview plays the exact audio that will end up in voice - what you hear is what your users hear.
  • Keep TTS short. Long announcements are slower to render and cost more per click. For greetings and queue prompts, shorter scripts are almost always better.
  • Iterate music prompts. Music generation is non-deterministic - the same prompt produces different clips on each run. If the first result isn't right, tweak the prompt (add "instrumental", "no vocals", "loopable") and regenerate.

Moderation and errors​

Both generators run every request through a safety check. If the prompt is rejected, you'll see a clear moderation message and no coins are deducted. Other errors the UI surfaces:

SituationWhat you'll see
Not enough AI CoinsThe generate button is disabled and a "Buy more coins" link appears.
Generation rejected (policy)"Your prompt was rejected by moderation" or similar - coins are not charged.
Too many concurrent requestsA "currently busy" notice - retry in a few seconds.
Generic failureA one-line error message. If it persists, reach out at scnx.app/help.
Brauchst du mehr Hilfe?
Findest du nicht, wonach du suchst? Unser Team hilft dir gerne.