An explainer video is a script with pictures attached. The script determines whether anyone understands the thing; the visuals determine whether they keep watching long enough to find out. Free tools handle the visual layer perfectly well, which means the money and the effort belong in the script and the voice.
And for software specifically, the best visual is not generated at all. It is a screen recording.
TL;DR
- Script first. 90 seconds is 220 to 240 spoken words, and that constraint is the useful part
- For software, record the screen. A real interface beats any generated abstraction of it
- Voice: ElevenLabs TTS at 13 credits per 1,000 characters, so a 90-second script is about 18 credits
- Visuals: generated stills with motion at 6 credits each, or plates at 24
- A full 90-second explainer: under 200 credits, about $2
Write the script to a word count
90 seconds is roughly 230 words at a comfortable pace. Writing to that number forces the decisions that make explainers work.
The structure that reliably lands:
- The problem, in one sentence the viewer would recognise as theirs. Not your product's category. Their Tuesday.
- Why the existing solution is annoying. One sentence, specific.
- What your thing does, stated plainly, no metaphor.
- One concrete example of it working. This is the part people remember, and the part most explainers skip.
- What to do next. One action.
Five beats, 230 words. Anything that does not fit one of the five is cut, and that rule alone fixes most explainers.
Visuals, in order of effectiveness
1. A screen recording of the real product. Free, specific, honest, and more persuasive than any abstraction. If the product is software, this is the answer and generation is a distraction.
2. Simple diagrams and text cards. A concept with three boxes and two arrows explains more than an illustration of a concept. Nano Banana Pro at 21 credits when the text in the image must be right, or build them in your editor for nothing.
3. Generated stills with motion added in the edit. 6 credits each on Seedream 5, then a slow push. Fills runtime, sets tone, costs almost nothing.
4. Generated video plates. Wan 3.0 at 480p, 24 credits per 5 seconds. Good for atmosphere between sections, rarely for explanation itself.
Note the ordering: the cheapest and least AI-heavy option is the most effective.
What it costs
Canon 8frame prices, a credit is $0.01 at pack rate. A 90-second explainer:
| Element | Route | Credits |
|---|---|---|
| Voiceover, ~1,400 characters | ElevenLabs TTS at 13 per 1k | ~18 |
| 12 illustrative stills | Seedream 5 at 6 | 72 |
| 3 diagrams with exact text | Nano Banana Pro at 21 | 63 |
| 2 atmosphere plates, 5s | Wan 3.0 at 480p | 48 |
| 30 seconds of music bed | Lyria 2 at 13 per 30s | 13 |
| Assembly | Merge Videos | 10 |
| Total | ~224 credits, about $2.24 |
Inside a Starter plan with most of the month left.
The voice decision
Synthetic narration is genuinely good for explainers, because the register is informational rather than emotional and that is where it performs best.
What to get right: punctuate for breath, since synthetic delivery follows punctuation literally. Write numbers as words where pronunciation matters. Listen to the whole track once, because product names and jargon are where it mispronounces, and those are your most important words.
When to use your own voice instead: when the explainer is about you or your company's judgement rather than a mechanism. A founder explaining why they built something should be audible as a person.
What not to do
Do not generate a presenter. An avatar explaining your product adds nothing and subtracts trust. Either narrate over visuals or film yourself.
Do not generate a fake interface. A plausible-looking UI that is not your UI misrepresents the product and disappoints on first use.
Do not stretch to three minutes. If the script does not fit 90 seconds, it usually contains two explainers.
Do not add captions as an afterthought. Most views are muted. See how to caption video for silent viewing.
FAQ
What is the best free explainer video maker? A screen recorder and a free editor, for software. Template tools work for the animated-diagram style.
How long should an explainer be? 90 seconds. 60 if you can.
Should I use an AI voice? For informational narration, yes. For anything about your judgement or your company, use your own.
How much does it cost? About $2.24 in credits for a full 90-second video.
Write 230 words, record the screen, narrate it. The 8frame canvas is free and unlimited, and generation is paid from $19/month.