How to Use Synthesia: A Job-by-Job Playbook

Synthesia earns its keep on a narrow set of jobs, and picking it for the wrong one is how most people end up let down. It is a specialist, not a generalist. It handles corporate training, onboarding, internal comms, and product explainers well, the kind of content you revise often and need to keep consistent across versions. Hand it persuasion, emotional storytelling, or anything that depends on showing a physical product or a live interface in motion, and it gives back much less; for those jobs, start from the Synthesia alternatives list or the HeyGen vs Synthesia comparison instead. The mechanics are simple, you pick an avatar, type your script scene by scene, choose a voice and language, and it renders the finished file with no camera or studio in the room, but the mechanics were never the constraint. The lane is. When people walk away unhappy, it is almost always because they brought it a job outside that lane, not because they hit a wall in the software.

A second thing shapes what you can build before any feature does. Synthesia moderates every video it renders, and its stock avatars are barred from endorsing products, sharing opinions, or covering medical, legal, financial, news, or political topics. If the video you have in mind is a review or a recommendation, a stock avatar will likely be rejected and you will need a personal or custom one instead. The sections below are organized around the jobs people bring to Synthesia, and each one flags where a rule like that changes how you build.

What Every Job Spends: Credits, Previews, and Render Time

Whatever you are making, it draws on the same pool of credits, so learn how they work first. Credits are the currency for everything, and the conversion is worth memorizing: 120 credits equals one finished minute, or two credits per second. Unused credits do not roll over to the next billing period, so there is no reason to hoard them. Drafting a script and clicking Preview cost nothing. Credits are only deducted when you click Generate, which means you can assemble and check an entire video for free and pay only at the render.

You can try all of this on a free Basic account at Synthesia, which you open with an email address. The vendor's own signup steps say no credit card is required for the free plan. Basic gives you 1,200 credits a month, roughly 10 minutes of video, plus 9 stock avatars, zero personal avatars, and a single editor seat. Every Basic video carries a "Made with Synthesia" watermark and cannot be downloaded at all, which is the hard wall: you can assemble scenes, preview the whole thing, and see exactly how the editor behaves, but the file never leaves Synthesia, so nothing you build here reaches an audience.

That "10 minutes" figure is a monthly credit allowance, not a length limit. A single Synthesia video can run up to 150 scenes, each up to 5 minutes, for a total of 4 hours. Your credit bucket is what caps you, not the editor. Synthesia runs in your browser, with Chrome and Edge listed as the supported ones, and the editor resembles PowerPoint: a toolbar across the top, a scene list down the left, the canvas in the center, a script box along the bottom, and an inspector panel on the right for whatever element you have selected. When a video is ready, Preview plays the whole thing at no credit cost. Rendering is not instant; the vendor describes it as taking a few minutes, so lock your script before you start.

Job One: Training and Onboarding You Rebuild Every Quarter

This is the job Synthesia fits best, because its cheapest trick rewards content you revise on a schedule. Re-generating an edited video only charges for the seconds you changed, and duplicating a video is free. A policy walkthrough or an onboarding module you tweak each quarter costs you minutes of edits rather than a full rebuild.

Start from the dashboard by clicking "New video" in the top-right corner. Synthesia offers four ways in. You can pick a Template and filter the library by tag, start from a Blank canvas for full control, or choose Import PowerPoint to turn a deck into scenes. The PowerPoint route accepts .pptx files only, up to 1 GB and 150 slides, and on import it asks whether to convert your speaker notes into scene scripts, which removes the most tedious part of setup when your source is already a deck. The fourth option is the AI Video Assistant, available on Starter and above but not on the free plan: you feed it a prompt describing your topic, audience, and goal, optionally attaching PDFs, Word files, or URLs, and it builds a storyboard of scenes with draft scripts you refine through a chat interface before clicking "Continue in editor." Because only the final Generate step charges credits, the Assistant costs nothing to use.

Inside the editor, the rule is one scene equals one script block. Break your content into scenes with one idea each rather than one long monologue. Avatar realism holds up on short clips but goes stiff and robotic on takes past roughly five minutes, so scene-based pacing keeps the delivery natural as a side effect. For course-style material, Interactivity (Creator and above) adds clickable calls-to-action, quizzes, and branching paths. Two limits to keep in mind: any scene with interactivity shows a hand icon in the scene list and cannot be exported as an MP4, and SCORM export for a learning system is Enterprise-only.

Job Two: Product and Process Explainers Built From a Screen Recording

When the point is to show how something works on screen, the AI Screen Recorder does the heavy lifting. It is a Chrome or Edge extension that captures your screen, transcribes the voiceover into an editable script, splits it into scenes, and syncs zoom and pan effects to your words. You can also upload an existing MP4, capped at 500 MB and 30 minutes, and Synthesia transcribes and scene-splits that too. Record any software walkthrough in short chunks with slow mouse movement so it stays in sync with the narration.

To dress a scene beyond the avatar, the toolbar adds Text, Shapes, Motion graphics, Media, Captions, Interactivity, and Recordings. The AI Playground tab generates video clips and images from text prompts using integrated Google Veo and OpenAI Sora models, useful for B-roll when a talking head alone feels flat. Brand kits that lock your colors, fonts, and logo as defaults are an Enterprise feature, while aspect ratio for everyone lives in the Video Actions menu. Keep the tool's weak spot in mind here: it does not show a live interface in motion the way a real screen capture does, so lean on the screen recorder and uploaded footage for the parts that have to be seen happening.

Job Three: Making the Narration Sound Intentional

Regardless of the job, the voice work happens in the script box, and a few controls separate a competent read from a robotic one. Type or paste your text, then use the speaker pill on the left of the script to assign which avatar reads it, and click the right side of that pill to change the voice or language. The editor's picker offers 140+ languages and 1000+ AI voices, and you can preview any voice with the play button at no cost.

Fine control lives in this box too. Hover the speaker pill for a speed slider that runs from 0.8x for slower delivery up to 1.2x for faster, and right-click the script for a context menu that inserts pauses from 0.1 seconds to 99 seconds or opens the Pronunciation tool. That last one is the one you want for brand names, acronyms, and jargon: type the phonetic spelling, preview the sentence, and Synthesia stops mangling your product name. The default text-to-speech sounds competent but recognizably synthetic, so short sentences and contractions read more naturally than dense paragraphs.

Job Four: One Video, Several Markets

If the same content needs to reach audiences in different languages, do not rebuild each version by hand. AI Dubbing, available on paid self-serve plans, re-voices a finished video across 70+ languages with lip sync (140+ on Enterprise), and Enterprise adds 1-Click Translation into 80+ languages that also translates on-canvas text. Bear in mind the dub is synthesized rather than voiced by a native speaker in the booth, so route it to internal training or broad marketing and put a fluent human on a review pass before anything high-stakes goes out.

Job Five: Reviews, Endorsements, and Anything With an Opinion

This is where Synthesia's rules come down hardest, and it catches affiliate creators off guard. Stock avatars may only carry neutral, fact-based content. They cannot endorse a product, deliver an opinion, represent an organization, reference real people or companies, or touch medical, legal, financial, news, or political subjects. A script that sails through on a personal avatar can be rejected on a stock one, so if your video recommends or reviews anything, budget for a personal or Studio avatar from the start rather than discovering the block at render time.

Putting your own face on screen means a Personal Avatar, and the count is gated hard: zero on Basic, 3 on Starter, 5 on Creator, unlimited on Enterprise. Two creation methods exist. The video method wants a single continuous take of 1 to 5 minutes with only one person in frame, and it processes in about one business day. The photo method, added in Synthesia 3.0, builds an avatar from a single image in minutes. Both require consent recorded live with an on-screen passcode, you cannot upload a pre-recorded consent clip, and you must be 18 or older. There is also a Studio Avatar add-on built on the Express-1 model at $1,000 a year, available only to annual subscribers, with up to 10 days of processing. Cloning your actual voice is a separate feature that, per Synthesia's help center, sits on the Enterprise plan: it records your consent live, has you read a few paragraphs, and produces a clone in under 15 minutes that can then speak across dozens of languages.

When you do use stock avatars for permitted content, the counts climb by tier: 9 on Basic, 125+ on Starter, 180+ on Creator, and 240+ on Enterprise. Two families matter. Express-1 is built from filmed footage of real actors, and Express-2 is the current generative engine that adds natural hand and body gestures, outputs 1080p at 30fps, and lets you switch camera angles per scene. Older "Legacy" and "Expressive" avatars are being phased out, so skip them. On the canvas, framing presets set Fullscreen, Circle, or Box, and multi-camera avatars are flagged with a special icon and support different angles. One more reason to plan avatars carefully around moderation: appeals are paid-plan-only, so free users cannot contest a rejection at all, and even paid users get a single appeal reviewed within 48 hours with a final answer. Never build a publishing deadline around a render that has not cleared moderation yet.

What Each Plan Actually Gates, and What It Costs

Read the pricing carefully before you commit, because the advertised number is not the monthly number. The $18 Starter and $64 Creator figures require a full year paid upfront; month-to-month costs $29 and $89. Purchases are non-refundable per Synthesia's Customer Terms, and cancelling only stops the next renewal rather than refunding unused time, so choose a billing cycle deliberately. The tiers also gate capability, not just avatar counts: the AI Video Assistant starts at Starter, Interactivity at Creator, API access at Creator, and SCORM and SSO at Enterprise. Resolution does not scale with price. Every tier, including the paid ones, renders at 1080p, so no plan unlocks 4K.

Getting the Finished Video Out and In Front of People

To export, open the three-dots menu and choose "Download." Synthesia offers .mp4 at Full HD 1080p, .wav for audio, .srt and .vtt for captions, and .xliff for translation. Basic and trial users cannot download at all. If a video contains interactive elements it will not export to MP4, and the workaround is to duplicate it, strip the interactivity, and regenerate a clean version. Sharing lives under the "Share" button, where you can create a public link, grab an embed code, protect a video with a password or require SSO, or attach a call-to-action. SCORM export and 1-Click Translation are Enterprise-only, and dedicated guides exist for pushing straight to YouTube and Vimeo.

FAQ

Can you use Synthesia without paying?

Sign up for the Basic plan with an email, no credit card needed for the free tier. It gives you 1,200 credits a month, about 10 minutes of video, plus 9 stock avatars and one editor seat. Because every video is watermarked and can't be downloaded, the free plan lets you assemble and preview a full video but never releases the file, so nothing you make on it ever leaves the platform.

How do you create a video in Synthesia?

Click "New video," then start from a template, a blank canvas, a PowerPoint import, or the AI Video Assistant. Add your avatar, type your script scene by scene in the box at the bottom, pick a voice and language, and hit Preview to check it for free. When it looks right, click Generate, which spends 120 credits per finished minute to render the final video.

Can you put your own face and voice in a Synthesia video?

Yes, through a Personal Avatar, though the count is gated by plan: zero on Basic, 3 on Starter, 5 on Creator. You create one from a 1-to-5-minute video take that processes in about a day, or from a single photo that generates in minutes, and both require consent recorded live with an on-screen passcode. Cloning your actual voice is a separate feature that sits on the Enterprise plan per Synthesia's help center.

Does Synthesia work in other languages?

Yes. The editor's voice picker offers 140+ languages and 1000+ voices. On paid self-serve plans, AI Dubbing re-voices a finished video across 70+ languages with lip sync (140+ on Enterprise), and Enterprise adds 1-Click Translation into 80+ languages that also handles on-screen text, so one video can become versions for several markets without reshooting.