Synthesia Fundamentals

Synthesia is an AI video platform that transforms written scripts into professional videos featuring realistic AI avatars. Founded in 2017, it is now trusted by over 50,000 companies including 70% of the Fortune 100. The core premise is simple: instead of recording a human presenter, you type what you want them to say, select an AI avatar, and Synthesia generates a video in minutes.

The platform eliminates the traditional production stack. You do not need a camera crew, a studio, a teleprompter, or a professional presenter. You do not need to schedule recording sessions or manage reshoots when content changes. When a product update requires a script change, you update the text and regenerate. The entire video is updated in minutes rather than days.

Synthesia operates entirely in the browser. There is no software to install. You log in at app.synthesia.io, create a new video, write or paste your script, and pick an avatar. The AI handles lip-sync, voice delivery, and facial expressions automatically. The result is a polished talking-head video that most viewers cannot distinguish from a real recording.

The platform is designed for business use cases: onboarding videos, product training, compliance modules, internal communications, and customer education. It is not a general-purpose creative video tool. Its strength is producing consistent, scalable, on-brand training and communication content at a fraction of traditional production cost.

The Synthesia editor is organized around a slide-based timeline. Each scene in your video corresponds to a slide. You add slides, write the script for each slide, set the avatar and background, and arrange them in order. The left panel shows your scene list. The center canvas shows the active scene. The right panel shows properties for the selected element.

Synthesia offers 230+ stock avatars across genders, ethnicities, ages, and presentation styles. Avatars are organized into categories: business formal, casual, expressive, and character. When choosing an avatar consider your audience and content tone. A compliance training module for a financial services firm calls for a professional formal avatar. An onboarding video for a tech startup might use a casual, friendly presenter.

Each avatar comes with a default voice, but you can mix and match. Select any avatar and then choose from hundreds of AI voices in your target language. You can also use voice cloning if you want your own voice on a different avatar.

Background options include solid colors, gradients, office environments, abstract scenes, and custom uploaded images. For most corporate training, a clean neutral background keeps the focus on the content. You can also add your company logo, insert text overlays, embed images, and include screen recordings within the same slide.

The AI script assistant is accessible from within the editor. Click the magic wand icon and describe what you want. The assistant drafts a scene-by-scene script that you can edit before generating.

Synthesia 3.0 introduced a deeply integrated AI script assistant that can generate entire video drafts from a single prompt. Describe your topic, target audience, tone, and desired length. The assistant creates a multi-scene script with suggested slide structure, avatar placement, and supporting visuals.

The assistant uses context about your workspace, brand kit, and previous videos to maintain consistency. If you have established a brand voice, the assistant learns from it. You can instruct it to match a formal tone, use simpler language for non-technical audiences, or mirror the style of an existing video in your library.

Beyond full-draft generation, the assistant supports targeted editing. Highlight a specific sentence and ask it to make it more concise, more engaging, or more technical. It will rewrite just that portion without touching the rest of your script. This is useful when you need to simplify a compliance paragraph or sharpen a call to action.

For teams producing high volumes of content, the AI can batch-generate script variations. If you need the same core message adapted for five different regional audiences, describe the variations and the assistant generates all five in a single workflow. You then review, adjust tone or terminology, and generate the videos.

The script assistant also supports importing source documents. Paste a PDF URL, a Word document, or raw text and it extracts the key information and structures it into a video script. This is the fastest way to convert existing training materials into AI video content.

Once your script is written and your scenes are configured, generating a video is a single click. Synthesia queues the render and sends an email notification when it is complete, typically within a few minutes for short videos. Longer videos with many scenes may take 15-30 minutes.

After rendering, the video appears in your library with a preview player. Review it carefully before sharing. Check lip-sync timing, voice naturalness, scene transitions, and text overlay positioning. If anything needs adjustment, return to the editor, make the change, and re-render. Because you are editing a script rather than video footage, revisions cost minutes not hours.

Publishing options include: a shareable link (anyone with the link can view), an embed code (paste into your LMS, intranet, or website), a direct download as MP4, and an export as a SCORM package for LMS tracking. The shareable link gives access to Synthesia's built-in video player which includes view analytics, chapter navigation, and language selection if you have translations.

For access control, enterprise plans let you set videos to require viewer login, restrict to specific email domains, or protect with a password. This is essential for internal training content that should not be publicly accessible.

Organizing your video library with folders and tags from the start saves significant time as your library grows. Create folders by department, project, or content type. Add tags like "compliance," "onboarding," or "product" to enable quick search across hundreds of videos.