Skip to content
DrömlikDocs
sv

Generate AI voice prompts

Last updated: 2026-06-01

Generate a lifelike audio prompt from text and use it anywhere Drömlik plays audio: IVRs, queues, ring groups, AI Receptionist greetings, or music on hold.

Steps

  1. 1

    Open System Preferences → Prompts → AI Text-to-Speech

    The library shows every TTS prompt already generated on this tenant.

  2. 2

    Click Add

    Give the prompt a clear name (e.g. "Sales queue — overflow EN"). You'll pick it from a list later, so make it findable.

  3. 3

    Type the text

    Up to 1,000 characters per prompt. Use punctuation — commas and periods control where the voice pauses.

  4. 4

    Pick language and voice

    Choose the language, then a standard or neural voice. Neural voices cost slightly more but sound noticeably more natural.

  5. 5

    Adjust speech rate (optional)

    Slow, normal, or fast. Stick with normal unless you have a reason.

  6. 6

    Preview

    Click Preview to hear it. If something sounds off — a brand name, a phone number — tweak the spelling (e.g. write "Drömlik" or "Drum-lik" depending on how it's pronouncing).

  7. 7

    Save

    The prompt is added to the library and becomes selectable wherever audio is configured.

Assigning a prompt

Open the IVR, queue, AI Receptionist, or Call Flow node you want to update. In the prompt picker, switch the source from Uploaded file to TTS library and pick your new prompt.

Pronouncing tricky words

  • For brand names or place names, try a phonetic spelling.
  • For phone numbers, separate digits with spaces or hyphens (e.g. 0 8 - 1 2 3 4 5 6) so each digit is read individually.
  • For URLs, write them in spoken form ("dromlik dot com").