Text-to-Speech for E-learning: A Guide for Course Creators
A practical guide to text-to-speech for e-learning: scripting lessons, consistent narration, technical terms, captions and budgeting credits for a course.
Text-to-speech is a practical way to narrate e-learning because course content changes. When a feature, policy or price is updated, you can edit one sentence in the script and regenerate that section in minutes, with the same voice, instead of re-recording or rebooking a narrator. This guide covers how to plan, script and produce course narration with AI voices, including consistency across modules, technical vocabulary, captions and budgeting.
Where text-to-speech fits in a course
AI narration works well for many kinds of learning content:
- Software tutorials and screen recordings, where the interface changes often.
- Compliance and policy training, which is updated regularly and needs a consistent, neutral tone.
- Onboarding modules for new staff or customers.
- Microlearning: short lessons, tips and quiz feedback clips.
- Multilingual versions of the same course. VoiceoverWise voices follow the language of the script and work in many languages, best in English.
Some content benefits from a human narrator: an emotional story, a founder's personal message or a flagship course where a familiar voice is part of the appeal. Our AI voiceover vs voice actor comparison discusses when each makes sense. Many teams use both.
Plan the course audio before you write
A little planning saves a lot of regeneration later.
- Choose one narrator voice for the course. Switching voices between lessons feels disjointed. Listen to all 11 voices and pick one that suits your audience.
- Choose a delivery style. Natural narrator suits most lessons; Professional suits compliance and finance; Warm & friendly suits onboarding and beginner material.
- Split lessons into sections. Generate one clip per slide or scene rather than one long file per module. Updating a single slide then means regenerating a single short clip.
- Name files consistently. For example, module-02-lesson-03-slide-04. It makes swapping updated clips painless.
Write scripts learners can follow
Learning narration needs to be even clearer than marketing narration, because learners are trying to absorb new information while watching the screen.
- Say what's coming. "In this lesson, you'll learn three things."
- Match the screen. Narrate what the learner sees, in the order they see it.
- One step per sentence in procedural content: "Open Settings. Select Users. Then choose Add user."
- Pause after key points. A
[pause](half a second) or[pause 2s]gives learners time to look at the screen or read a definition. Pauses go up to five seconds. - Recap at the end. One or two sentences that restate the key point.
A sample slide script:
To add a new user, open Settings. [pause] Select Users, then choose Add user. [pause] Enter their email address, for example [email protected]. [pause 2s] They'll receive an invitation straight away.
With smart reading, that email is read as "jordan at example dot com". If your source material is in slides or documents, the free TTS script cleaner removes bullets and formatting before you paste it in.
Handle technical terms once
Courses are full of product names, acronyms, jargon and people's names, and the same words appear in every lesson. Instead of fixing them each time, save them in the pronunciation dictionary. Once a word is saved, it's applied to every script automatically, so "Kubernetes", your product name or a lecturer's surname sound the same in lesson one and lesson forty.
Smart reading also handles numbers learners care about: prices, percentages, ISO dates such as 2026-10-12 (read as "October 12, 2026"), phone numbers digit by digit, and abbreviations like e.g. and etc. Check the free preview to see exactly what will be read. Our guide to fixing mispronunciations has respelling tips.
Produce at scale with bulk mode and the API
A full course can have hundreds of short narration clips. Bulk mode lets you upload a CSV or paste lines, up to 200 rows at a time, and get every audio file back in one ZIP. A simple approach is one row per slide, using the slide's file name as an identifier so clips drop straight into place.
If you build your own learning platform, the REST API lets you generate narration programmatically, for example whenever a lesson's text is updated. See the API documentation.
Download each clip as MP3 or WAV. MP3 is fine for most authoring tools; WAV is a better master if you edit audio in a video editor. Save your files locally: audio is available for 30 days, then deleted, while scripts stay in your history so you can regenerate later.
Captions and accessibility
Every clip comes with an SRT subtitle file timed per sentence. Timing is estimated from the audio, not word-level, so review it in your authoring tool, but it gives you a starting caption track for every slide without extra work. Captions help learners who are deaf or hard of hearing, learners studying in a second language and anyone watching without sound. Check the accessibility standards that apply to your organization; captions are often one part of them. SRT subtitles explained covers editing and uploading.
Budgeting credits for a course
Credits are based on characters: 1 credit per character, about 900 characters per minute of audio. To estimate a course:
| Course size | Approximate narration | Approximate credits |
|---|---|---|
| Short module | 10 minutes | About 9,000 |
| Small course | 1 hour | About 54,000 |
| Full course | 5 hours | About 270,000 |
Allow extra for revisions and regenerated sections. The free plan includes 10,000 credits every month (about 11 minutes), enough to pilot a module. For a full course, one-time packs start at $29 for 300,000 credits (about 5.5 hours), valid for 12 months, with no subscription. See the pricing page, and paste a lesson into the script timer to estimate length.
The easiest way to decide is to narrate one real lesson. Start free, pick a voice and generate your first module with captions included.