OndaLabs Text to Speech: Make Your First Voiceover

To make your first OndaLabs voiceover, sign up free, open text to speech, pick one of 30 voices, paste your script, add a style and optional expressive tags, then generate and download an MP3. The whole path works on the Free plan, which includes 2,000 credits a month, about 2 minutes of speech on Onda Lite.

This walkthrough follows the studio as it is labelled today. Button and field names are quoted as the app shows them.

Five steps to a first voiceover

Create a free account.
Open text to speech and choose a voice.
Write your script and check the counter.
Direct the delivery with a style and tags.
Generate, listen, and download the MP3.

Step 1: Create a free account

Sign up with Google, or with your email, first name, last name and a password. After sign-up the app sends you to an email verification page before the studio opens. If you visit the studio while signed out, you are sent to the login page.

The Free plan gives you 2,000 credits each month, about 2 minutes of speech. It always uses Onda Lite and includes expressive tags, dialogue, speech to text and 5 saved clips. You can sign up here at ondalabs.app, and compare the other plans on the pricing page.

Step 2: Open text to speech and pick a voice

Open text to speech. The default voice is Kore, labelled Firm. Press Preview voice to hear it before you commit, or choose another voice. All 30 voices speak every supported language, and the app detects the language of your script on its own.

A few voices that suit common jobs:

| Voice | Character | Suggested use | |---|---|---| | Kore | Firm (default) | Direct explainers | | Sulafat | Warm | Gentle narration and bedtime stories | | Charon | Informative | Lessons and tutorials | | Algenib | Gravelly | Characters with some grit |

The voice library can filter voices by Warm, Bright, Calm, Upbeat and Informative, and by gender. The suggested uses in this table are our own picks, not product labels.

Step 3: Write your script

The text box reads "Write something worth hearing. Type or paste your script here...". One generation accepts up to 5,000 characters. For longer work, you can split the script into up to 50 speech blocks with Add speech block, and each block can have its own style. All blocks in one generation still count toward the 5,000 limit.

Under the editor, a counter reads "N / 5,000 characters" next to the credit cost. Blocks are joined with a space before counting. Tags count as characters too, because they are part of the text. That means <laugh> costs 7 characters.

Step 4: Direct the delivery

Each script block has two chips. The Style chip takes a free-text direction of up to 200 characters, or one of 10 presets. The Expression chip opens 33 expressive tags in seven groups. Clicking a tag inserts it at the cursor, so you can place a laugh or a pause exactly where you want it. You can also type a tag by hand.

The 10 style presets and what each one sends with your text:

Friendly: warm and friendly, relaxed, small natural pauses
Narration: commanding storyteller, rich vocal modulation, dramatic pauses
Calm: calm and soothing, slow, soft and even
Excited: excited and high energy, bright and quick
Whisper: whispering quietly, close to the microphone
Newscaster: clear, confident newsroom delivery, steady pace
Promo: upbeat commercial voiceover, punchy, elongated vowels
Sarcastic: dry and sarcastic, deadpan timing
Sad: sad and subdued, quiet and slow
Angry: angry and tense, clipped words

The 33 tags are grouped as follows. Write each one inside its angle brackets, as shown in the table.

| Group | Tags | |---|---| | Pauses | <short pause>, <long pause> | | Laughter | <laugh>, <chuckle>, <giggle>, <snicker>, <cackle>, <snort> | | Breath | <breath>, <heavy breath>, <exhales>, <sigh>, <pant>, <yawn>, <phew> | | Emotion | <gasp>, <cry>, <sob>, <whimper>, <groan>, <moan> | | Loud | <cheer>, <shout>, <scream>, <shriek> | | Sounds | <cough>, <sneeze>, <throat-clearing>, <grunt>, <growl>, <hiss>, <tsk> | | Voice | <whispers> |

Use tags sparingly. A tag every sentence rarely sounds like a person talking.

Step 5: Choose a model, check the cost, and generate

Free accounts always generate with Onda Lite. Paid plans can pick Onda Flash as well. The credit cost is the character count multiplied by the model's rate, rounded up. Lite costs 1 credit per character and Flash costs 1.5.

The button reads "Generate speech". If your balance is too low for the job, the button becomes "Add credits" and opens a window that says you are out of credits.

Finished audio appears in a player with a download button. The download is an MP3 file named like ondalabs-<voice>-<timestamp>.mp3.

Listen before you download. Each new generation costs credits again, so change one thing at a time: the voice, the style or a single tag. A short first script keeps the cost low. A 300-character paragraph costs 300 credits on Onda Lite and 450 on Onda Flash, using the same formula as above.

Two built-in examples show the cost difference:

The "Ad voiceover" example is 140 characters. It costs 140 credits on Onda Lite and 210 on Onda Flash.
The "Bedtime storyteller" example is 232 characters, including two tags. It costs 232 credits on Onda Lite and 348 on Onda Flash.

Try a built-in example first

While the editor is empty, the studio shows ready-made examples. There are six speech scripts (Bedtime storyteller, Ad voiceover, Morning news, Meditation guide, Game character and Friendly tutor) and three dialogue examples. The samples are pre-recorded, so listening to them costs no credits.

Clicking an example loads its script into the editor. It is the quickest way to hear how a style direction and a tag sound before you write your own script.

How long will the script run?

Credits depend on characters, but the length of the audio depends on pace. The app estimates about 900 characters per minute on Onda Lite, based on about 15 characters of speech per second.

For a rough cross-check, ACX's guide to narration length says most performers narrate about 9,300 words per hour, which is about 155 words per minute. Multiply 155 by 5.8 and you get about 900. The two estimates agree if an average word plus its space takes roughly 5.8 characters. That is a rough check, not a rule. Real pacing varies with voice and script.

What to try next

Once you have one clip, try a two-voice scene in dialogue. The credits guide explains how a clip's cost adds up over a month. The Onda Lite vs Onda Flash comparison helps you choose a model for each script.

Frequently asked questions

Is OndaLabs text to speech free?

Yes, there is a Free plan with 2,000 credits a month, about 2 minutes of speech on Onda Lite. It includes expressive tags and dialogue. Onda Flash, image generation and commercial use need a paid plan, and the Free plan is for personal use and testing.

How many characters can I convert at once?

One generation takes up to 5,000 characters across all of its speech blocks. For longer work, split the script into several generations. Dialogue has the same 5,000 character limit per generation.

Can I download the audio as an MP3?

Yes. Finished audio appears in a player with a download button, and the download is an MP3 file. Use the player to listen first, then download the clip you want.

How do I add a pause, laugh or breath to AI speech?

Open the Expression chip and insert a tag such as <short pause>, <laugh> or <breath> at the cursor. Tags count toward your character total, so <laugh> costs 7 characters. You can also type a tag directly into the script.

Does OndaLabs speak languages other than English?

Yes. All 30 voices speak every supported language and detect the language of your script automatically. Dedicated text to speech pages exist for 37 languages.