Recordings, uploads & text-to-speech
Overview
Everywhere your hotline needs a sound — a menu greeting, a prompt, an instruction, a shiur — you pick it with the same Select Audio button. One modal, seven ways to get audio in: upload a file, reuse something you already have, record straight from your microphone, or have the system speak text for you.
This page covers that picker, bulk uploading, storage limits, and the text-to-speech tools. For managing recordings once they are in an audio player — archiving, transcripts, reordering — see Audio player (Library).
The audio selector
Click Select Audio on any audio field. Choose a tab, fill it in, and press Use This Audio. A Clear button next to a chosen file removes it.
You only fill in one tab. If more than one is filled, a chosen library recording wins, then dynamic text-to-speech, then generated speech, then an upload.
Upload New
Upload Audio File — MP3, WAV, OGG or M4A. Anything that is not already a WAV is converted automatically for you. This is the tab to use unless you have a reason not to.
Upload MP3
Upload MP3 File — an .mp3 kept exactly as it is, with no conversion, up to 200 MB. Use this when the file is already how you want it and you would rather not have it touched, or when it is too long for the converter.
Library Recordings
Reuse something already on your account.
- Filter by Library — narrow to one audio player, or leave it on All Libraries.
- Select Recording — searchable, showing the recording name and which player it came from. The 100 most recent are listed.
Record
Record from Microphone — record straight in your browser. Saved as a WAV. Useful for a quick greeting without leaving the page.
Generate TTS
Type words, get audio.
- Text to Speech — the text to be spoken.
- Voice — Alloy, Echo, Fable, Onyx, Nova or Shimmer. Echo is the default.
The audio is generated when you press Use This Audio and stored on your account like any other file.
Dynamic TTS
The same idea, but the words are built freshly on every single call, so the audio can say the caller's name or today's date.
- Dynamic script — your text with placeholders in it, for example:
Hi {first_name}, today is {date}. We will give you a call back soon. - Voice — the same six voices, Echo by default.
Available placeholders:
| Placeholder | What it becomes |
|---|---|
{first_name} |
The caller's first name, if they match a contact |
{last_name} |
Their last name |
{full_name} |
Their full name |
{phone_number} |
The number the caller is calling from |
{called_number} |
The number they dialed |
{date} |
Today's date |
{day} |
Today's day |
{time} |
The current time |
{account_name} |
Your account name from Account settings |
Your own contact fields become placeholders too. Anything you define under Contact fields in Account settings can be spoken as {its_key}, and the helper text under the script lists exactly which ones your account has.
Unknown placeholders are left as-is — if you type {nickname} and no such field exists, callers hear the braces. Check your spelling against the list in the helper text.
From URL
Audio URL — paste a link to an audio file and it is fetched and saved onto your account. Once imported it is yours; the original link does not need to stay up.
Uploading many files at once
Where you are adding recordings to an audio player, there is an Upload many button (on the My Hotline page the same thing appears as "Upload recordings into …").
Drag in as many audio files as you like. Each file becomes its own recording, named after the file with the extension stripped — so 01 Bereishis.mp3 becomes a recording called 01 Bereishis.
This is the fast way to load a whole set of shiurim. Name the files properly before you drag them in and you will not have to rename anything afterwards.
Storage
Every file you upload, every message a caller leaves, and every piece of generated speech uses storage. The base plan includes 500 MB, and extra space is $2 per GB per month on Plans & Add-ons. Roughly 15 to 20 hours of recordings fit in a GB.
Watch your usage on the Storage Used tile of the Statistics page.
When storage is full
What happens depends on the choice you made under When storage fills up on the Plans page:
- Auto-grow on — we add 1 GB for $2 a month automatically and email you. Uploads never stop.
- Auto-grow off (the default) — uploads are refused with the message "Your storage is full — free some space or add more on the Plans page", and we email you.
If you hit that message, either delete or archive recordings you no longer need, or add storage on Plans & Add-ons.
How generated speech is reused
Generated speech is cached per text and voice. Ask for the same words in the same voice a second time and the existing file is reused instantly, with nothing new generated and nothing extra stored.
Change one word, or switch voice, and a new file is made. So there is no cost to reusing a standard phrase all over your hotline, but there is a small cost to endlessly tweaking wording.
The generators on My Hotline
The My Hotline page has two builders that write scripts for you from what you have already set up.
Generate the menu recording
"A spoken menu built from your keys — pick which to announce, polish the words, choose a voice."
- Which keys to announce — tick boxes, one per key on the current menu, each showing the phrase it would contribute. Keys behind an access check start unticked, because announcing a restricted area should be a deliberate decision.
- The menu script — the draft, assembled from the ticked keys. Edit it freely; ticking keys again rewrites it.
- Voice — the same six voices, Echo by default.
Generate & use it creates the audio and sets it as that menu's greeting straight away. Press play above the keypad to hear it.
Spoken instructions for listeners
Inside the In-call keys panel. It drafts the announcement of your listening controls from the keys you actually have, plus your go-back key.
- The wording — editable draft.
- Voice — the six voices.
- Save the audio into — a new "Instructions" audio player, or an existing one.
Full details in Account settings & keypad controls.
Transcription
Any recording in an audio player can be transcribed into text you can read, search and email. Select it and use Transcribe, or Transcribe Selected for a batch, then read the result with View Transcript (text, summary, keywords, word count).
Transcription is charged at $0.0001 per YiddishLabs credit against your balance — roughly a few cents per minute of audio. If you put your own YiddishLabs API key into Account settings, YiddishLabs bills you directly and nothing is charged here.
The full workflow, the transcript filters, and emailing transcripts are covered in Audio player (Library).
Tips
- Record or upload once, reuse everywhere. Put shared prompts in one audio player and pick them from the Library Recordings tab rather than uploading the same file repeatedly.
- Use Generate TTS for the plumbing — "please hold", "returning to the main menu" — and a real recorded voice for anything your callers will hear often.
- Use Dynamic TTS for anything personal. A static recording cannot say a caller's name; a dynamic script can.
- Name files before bulk uploading. The filename becomes the recording name, so a tidy folder becomes a tidy audio player.
- Test on the phone, not the browser. Audio that sounds fine on a laptop can be too quiet down a phone line. Use your test PIN to check.