How to Get Started with Synthesia (2026 Guide)
A practical step-by-step guide to setting up Synthesia properly — including the one mistake most teams make.
Some links below are affiliate links — we may earn a commission at no cost to you. This never affects our assessments.
Synthesia turns a text script into a video with a realistic AI avatar presenting it, in 140+ languages, without a camera, actor, or studio. Here is how to actually get a usable video out of it.
Step 1: Write the script first, not in the tool
Write and edit your script in a separate document before pasting it in — the avatar reads exactly what you give it, so typos, awkward phrasing, and run-on sentences all come through in the delivery. Short sentences read more naturally than long ones.
Step 2: Choose an avatar and voice that fit the tone
Synthesia offers a library of stock avatars plus the option to create a custom avatar of yourself or a real presenter. For internal training content, a stock avatar is usually fine; for anything client-facing, a custom avatar matching your brand reads better.
Step 3: Break long content into scenes
Rather than one long unbroken script, split content into scenes with on-screen text, images, or slides alongside the avatar — this keeps viewers engaged and makes the video easier to edit later if one section needs updating.
Step 4: Review the AI pronunciation before finalizing
Technical terms, product names, and acronyms sometimes get mispronounced — preview the full video and use Synthesia's phonetic spelling override for anything that comes out wrong before publishing.
Looking for a software agency?
Take our free 3-minute matching quiz and get a personalised agency shortlist.