Audio files and text-to-speech

Every greeting, hold message, and IVR prompt on your dialer lives in one place: the Audios tab. Upload a recording from your computer, or type a script and let text-to-speech turn it into a phone-ready prompt in seconds. This walkthrough covers both, plus how to wire a prompt into your call flows. About five minutes.

Step 1 — Open the Audios tab

1. Click Servers in the left sidebar.
2. Click your server’s name in the list.
3. A row of tabs appears under "VICIdial Management & Reporting". Click Audios.

You are now looking at every audio prompt on your server: filename, format, size, length, when it was uploaded, and whether it is in the right format for the dialer (the Valid column — a green "good" is what you want). The list refreshes itself every 30 seconds, or click Refresh for an instant update.

Step 2 — Upload an audio file

Have a recording already? A studio greeting, a voicemail you exported, an MP3 someone emailed you — all fine.

1. Click Upload new audio, top-right of the table.
2. Next to File, click Choose File and pick the recording. WAV, MP3, and GSM files are accepted, up to 10 MB.
3. Under Save as, give it a name you will recognize later — for example main-greeting.wav. Letters, numbers, underscores and dashes only.
4. If a file with that name already exists, an Overwrite existing file with this name checkbox appears. Tick it only if you mean to replace the old prompt everywhere it is used.
5. Click Upload. The new row appears in the table within a second or two.

One thing worth knowing: whatever you upload, your server automatically converts it to the exact format VICIdial plays on calls — 8 kHz mono WAV, the telephone standard. Upload an MP3 and it still lands as a .wav file. This conversion is what keeps your prompts from playing as silence, which is what happens to wrong-format files dropped on a server by other means.

Step 3 — Generate a prompt from text (TTS)

No microphone? Type the script and we synthesize a natural-sounding voice for you.

1. Click Generate with TTS, just left of Upload new audio.
2. Under Voice, open the dropdown and pick a voice. Click the little play button next to it to hear a sample — samples are always free.
3. Under Text, type or paste your script. The counter in the corner shows how much room you have: up to 2,000 characters, roughly two minutes of speech. Write it exactly as you want it spoken; punctuation is how you add pauses.
4. Watch the Cost tile as you type — it shows the exact price before you commit, usually a cent or less for a greeting. If your plan includes free monthly characters, the tile says so and the cost shows as free. Next to it, the Wallet balance tile shows what you have available.
5. Click Preview. This is the moment the small charge happens (or your free characters are used) — the voice is generated for real so you can hear the finished result.
6. Listen in the player that appears. Not happy? Click Discard and tweak your text or voice. The charge is not refunded — the audio was really generated — so audition voices with the free sample first.
7. Happy? Click Save audio, type a name (we add .wav for you), and click Save audio again. The prompt lands in your audio library, exactly like an upload.

If your wallet can’t cover the cost, the Preview button swaps to an Add funds to wallet link. And if the generation itself fails after charging, the charge is automatically refunded — you will see it as an "Adjustment" entry in your wallet history.

Step 4 — Listen, download, rename, delete

Click any filename in the table to open the player: format, size, length, and validity up top, a play bar in the middle, and a Download button if you need a copy on your computer.

• The pencil icon renames a file. Careful: everything that uses this prompt references it by name, so renaming breaks those references — update them right after.
• The trash icon deletes (owner accounts only). You will be asked to type the exact filename to confirm. Any in-group, call menu, or schedule still pointing at the file will simply stop playing that prompt.

Step 5 — Use the prompt in your call flows

Prompts do nothing until something references them. Where to wire them in:

An IVR (call menu)

1. Click the Inbounds tab, then the Call Menus pill.
2. Click the pencil icon on your menu’s row.
3. On Basic Details, use the dropdowns: Greeting prompt (plays when the caller reaches the menu), Timeout prompt (plays if they press nothing), and Invalid prompt (plays on a wrong key). Each dropdown has a play button so you can listen before choosing.
4. Click Save changes.

An in-group (queue)

1. Click the Inbounds tab, then the In-groups pill.
2. Click the pencil icon on your in-group’s row.
3. Open the Advanced tab and type your filename into Group announce filename — that plays to callers as they enter the queue. The Filters & Alerts tab has Second alert filename and Third alert filename for agent wake-up alerts.
4. These fields are free text, not dropdowns — copy the exact filename from the Audios tab.
5. Click Save changes.

After-hours messages

1. Click the Schedules tab.
2. On the Call Times pill, click the pencil icon on a call time.
3. Set Default after-hours audio, or a per-day After-hours audio next to each day’s hours. Holidays (the Holidays pill) have their own After-hours audio field.
4. Click Save changes.

Live calls

Agents can also trigger prompts by hand during a call — that is the soundboard. See the Soundboards article for that setup.

What about outbound campaigns? There is deliberately no audio dropdown on the campaign editor — answering-machine and safe-harbor messages on outbound campaigns are wired through dialplan extensions, which is advanced territory. If you need one, ask support and we will set it up with you.

Good to know

• Everything lands as 8 kHz mono WAV, no matter what you upload or generate — that is the one format VICIdial plays reliably. If you ever see "bad" in the Valid column, re-upload through the Audios tab or regenerate with TTS.
• TTS charges at Preview, not at Save. Discard is a clean-up, not a refund — use the free voice samples to audition first.
• A TTS preview expires after about an hour. If you wait too long to save you will get "Preview expired" and need to generate again (and pay again) — save promptly.
• If a name collides when saving a TTS prompt, you keep your preview and can retry with a different name — no second charge.
• Filenames are the glue. Call menus, in-groups, and schedules all reference prompts by exact filename; renaming or deleting a file does not update those references for you.
• TTS is capped at 2,000 characters per generation. Longer script? Split it into separate prompts — call menus want short snippets anyway.
• Uploads appear in the table instantly and reach the dialer’s playback folder within about a minute — no restart, no downtime.
• Only upload recordings you have the rights to use, and keep the original files somewhere safe.