Synthesia 2.0: Create Training Videos in 45 Languages Without a Camera Crew

🎯 Why This Matters

The obvious way to make a multilingual training video is to record it once, then pay translators and voice actors per language, then re-edit each version. That's weeks of coordination and thousands of dollars per language. Synthesia collapses that into one master video plus a translation pass — and because the avatar is synthetic, the lip sync actually matches each language instead of looking dubbed. The other reason it wins: you can update the script once and regenerate every language in minutes, which is impossible with human-recorded video.

📖 Overview

A product manager at a 40-person SaaS company shipped a full onboarding curriculum in 11 languages in a single afternoon — no camera, no studio, no voice actors. The trick isn't just picking an avatar and typing a script. It's a specific workflow that uses Synthesia's translation and voice-cloning pipeline to produce one master video, then fan it out across every language your team actually needs, with the avatar's lip movements matching each language.

You need a paid Synthesia plan (Starter or higher) for the multilingual features and custom avatar slots. A free account will let you test the editor but won't give you the language fan-out or the voice cloning. You also need a script — ideally one you already have in a doc, wiki, or LMS.

By the end of this workflow you'll have a single 3–5 minute training video that exists in every target language, with the same presenter, the same slides, and lip-synced delivery in each. You'll also have a reusable template so the next video takes 20 minutes instead of a day.

Time estimate: 60–90 minutes for the first video, 15–25 minutes for every one after that. Output format: MP4 files per language, plus a shareable link and an embed code for your LMS or intranet. Captions and a transcript are generated automatically per language.

🛠️ Step-by-Step Guide

Step 1

Log into Synthesia and click 'Create video' from the dashboard. Choose 'Start from scratch' rather than a template for your first run — you want to see every setting. [Screenshot: Synthesia dashboard with the Create video button highlighted and the template gallery visible below it.]

Step 2

In the editor, select a stock avatar from the avatar panel on the right. Pick one that matches the tone of your training — a professional presenter for compliance content, a friendlier avatar for onboarding. Set the avatar's position to 'Full body' if your plan supports it, otherwise 'Half body' is fine for training. [Screenshot: avatar selection panel showing stock avatars with the Full body / Half body toggle.]

Step 3

Paste your script into the script box for the first scene. Break it into scenes of 30–60 seconds each — one scene per slide or per concept. Use the 'Add scene' button between concepts. Keep sentences under 20 words; Synthesia's TTS handles short sentences more naturally and it makes translation cleaner. [Screenshot: script editor with multiple scenes listed and the Add scene button visible.]

Step 4

Open the Voice settings and either pick a stock voice in your source language or, if you have a custom voice clone set up, select it. For custom voice cloning, go to Settings > Voices > Create voice and upload 2–3 minutes of clean audio — do this once and reuse it forever. Set the speaking pace to 'Normal' (not 'Fast') — fast pacing degrades badly in translation. [Screenshot: Voice settings panel with stock voices and the custom voice option.]

Step 5

Add your slides or screen recordings as background elements per scene. Use the 'Media' panel to upload PNGs, JPGs, or MP4s. Keep text on slides minimal — the avatar is speaking, so the slide should support, not duplicate. Set slide transitions to 'Fade' at 0.3s for a clean look. [Screenshot: Media panel with uploaded slides and the transition settings.]

Step 6

Once the source-language video looks right, click 'Translate' in the top toolbar. Select every target language you need from the list. Synthesia will generate a translated script, translated on-screen text, and a lip-synced avatar for each language. Review each translated script for tone — machine translation still mangles idioms, so spot-check the first 30 seconds of each language. [Screenshot: Translate dialog with language checkboxes and the Generate button.]

Step 7

For each language, open the translated version and check: (a) the avatar's lip sync, (b) any on-screen text that should have been translated but wasn't (logos and product names usually shouldn't be), and (c) the pacing. Fix any script issues by editing the translated script directly — Synthesia re-renders the avatar automatically. [Screenshot: translated video preview with the script editor open.]

Step 8

Click 'Export' and choose MP4 for each language. Also download the SRT caption file for each — your LMS will need it. If you want a single shareable link that switches languages, use the 'Share' option and enable the language selector. Copy the embed code if you're dropping it into an LMS like Docebo, TalentLMS, or a Notion page. [Screenshot: Export dialog with MP4, SRT, and share-link options.]