Synthesia

Synthesia is an AI video platform focused on business and training content. You pick an AI avatar, write or paste a script, and it produces a presenter video with matching speech in over a hundred languages. It is widely used for training, onboarding, internal communications and product explainers, and it also offers custom avatars, templates and, on higher plans, an API.

Quick facts

• Made by: Synthesia, a company based in London, founded in 2017
• Free plan: Yes, with a small monthly video allowance, a limited set of avatars and a watermark
• Paid plans: Starter and Creator, plus custom Enterprise
• Works on: Web, with API access on higher plans
• Pricing last checked: October 2026, using several recent published sources. Free-tier minutes and annual prices are reported differently across sources, so confirm on Synthesia’s official pricing page.

What Synthesia is best at

• Training and onboarding videos: consistent presenter videos without filming.
• Many languages: narration in well over a hundred languages and accents.
• Custom avatars: digital versions of yourself or team members, with consent, on paid plans.
• Templates and editing tools: scenes, screen recording, captions and background music.
• Enterprise features: SSO, SCORM export and dedicated support at the top tier.

Plans and pricing

Minutes are often shown per year on annual plans, for example about 120 a year on Starter and 360 on Creator, which equals roughly 10 and 30 a month. Minutes do not roll over, so the cost per finished minute can be high. Enterprise deals are quoted individually, and third-party sources report they often land in the low five figures a year. Confirm all figures on Synthesia’s official page.

Best for

• Learning and development teams producing training videos at scale.
• Businesses translating the same video into many languages.
• Teams that want a consistent on-screen presenter.

Less suited to

• Long or high-volume video on the lower plans, because minutes are capped.
• Content that needs real emotion from real people, where avatars may feel artificial.
• Anyone who wants to use someone’s likeness or voice without their clear permission.

How to get started

1. Sign up for the free plan and make a short test video with a stock avatar.
2. Write the script to be spoken aloud, with short sentences and one message per video.
3. Check how many minutes you need each month before choosing a plan.
4. Try a translated version of one video to test the languages you need.
5. Get written consent before creating any avatar or cloned voice of a real person.

Prompt ideas to copy

Write a 90-second onboarding video script for new [role] employees. Cover who we are, what to do in the first week and who to ask for help, in a friendly, clear tone.

Turn this policy document into a three-minute training script with a short quiz question at the end of each section.

Rewrite this script so it sounds natural when spoken aloud, with short sentences, plain words and no jargon.

Translate this script into [language] for an internal audience, keeping the tone professional and flagging any terms that do not translate well.

Limitations and things to watch

• Monthly minutes are limited and do not roll over.
• Avatar videos can look less natural than real footage.
• Sources disagree on free-tier minutes and some annual prices.
• Follow local rules on consent and on labelling AI-generated video.

Alternatives

Compare HeyGen for avatar and translation video, Descript for editing recorded video, and ElevenLabs for voices.