AI presenter videos from a script, built for training and explainers.
VideoMarketingEducation
AI presenter videos from a script, built for training and explainers. Write a script, choose a stock or personal AI avatar, pick language and voice, then generate a presenter-led video. Minutes and credits are plan-capped; dubbing and voice features draw from the same credit pool. Synthesia supports many languages and voices (exact counts vary by plan), stock and personal avatars, optional voice cloning with a custom personal avatar, and API access on Creator and Enterprise plans.
One topic per video or chapter. Lead each scene with a clear spoken line for the avatar, then support with on-screen bullets or b-roll—not long paragraphs the presenter reads verbatim without breaks.
What to provide
Details that improve the result
Example prompt
3-scene onboarding video for new sales hires. Scene 1: welcome and goals (45s). Scene 2: CRM login steps—script must say the product name 'Northwind CRM' exactly. Scene 3: quiz reminder. Professional tone, medium-paced voice, 16:9, stock avatar in business casual.
If the first output is not good
Common mistakes
Open with outcome (“By the end you will…”), then one step per scene. Keep spoken lines concise; put detail on screen when possible.
What to provide
Details that improve the result
Example prompt
90-second explainer: export a report from Analytics. Scenes: (1) navigate to Reports, (2) choose date range, (3) download CSV. Script uses exact button labels. Friendly explainer tone, 16:9, light office background.
If the first output is not good
Common mistakes
Finalize the primary-language video first, then dub or recreate scenes per language. AI dubbing draws from the shared credit pool—budget languages up front.
What to provide
Details that improve the result
Example prompt
Base English training video (3 scenes, 4 minutes total). Dub to European Spanish and French for EU staff. Keep on-screen product name in English; translate spoken instructions only.
If the first output is not good
Common mistakes
Pick a stock avatar that fits audience and tone, then write to the avatar’s natural pacing. Stock avatar counts vary by plan—confirm availability before scripting multiple cast members.
What to provide
Details that improve the result
Example prompt
Internal policy update, 2 minutes. Calm professional stock avatar, neutral studio background, navy lower-third bar. Script is plain language, no jokes, 16:9.
If the first output is not good
Common mistakes
Voice cloning can be paired with a custom personal avatar on supported plans. Record clean sample audio if required, then test short lines before generating long training content.
What to provide
Details that improve the result
Example prompt
CEO update video, 60 seconds, using approved personal avatar and cloned voice. Script: three short paragraphs, contractions allowed, no legal disclaimers beyond the provided compliance sentence.
If the first output is not good
Common mistakes
Lock design in the template; swap only approved variables (name, role, metric, language). Batch similar videos instead of one-off edits when content repeats.
What to provide
Details that improve the result
Example prompt
exploreMonthly manager update template: slots {manager_name}, {team_win}, {metric_value}. Generate 12 versions from a spreadsheet. Fixed closing compliance line must appear verbatim in every video.If the first output is not good
Common mistakes
API access is included on Creator and Enterprise plans. Treat scripts and template variables as your contract; validate render status before publishing URLs downstream.
What to provide
Details that improve the result
Example prompt
POST generate from template_id=onboarding_v3 with variables { employee_name, start_date } and avatar_id=stock_042, voice=en-US. Poll until complete, then save MP4 URL to HRIS record.If the first output is not good
Common mistakes
Synthesia is script-first: you write what the avatar should say, then shape scenes, voice, and visuals. Prompt-like planning means clear spoken lines, scene boundaries, pronunciation hints, and on-screen support—not open-ended chat. Budget minutes and credits, especially for dubbing and batch runs.
Video: 2-scene GDPR refresher, 16:9, professional tone. Scene 1 (40s): define personal data—use exact phrase 'personal data' twice. Scene 2 (50s): three employee actions as numbered spoken lines. Avatar: stock business casual. Voice: en-GB, medium pace. On-screen: max 5-word bullets per scene.
Short sentences, contractions where appropriate, and pauses between ideas sound more natural in presenter videos.
Example
Say: 'Open Settings. Choose Privacy. Turn on two-factor auth.' Not: 'Users should navigate to the Settings area in order to configure privacy-related authentication enhancements.'
Scene breaks map to regeneration boundaries and keep credits focused when one section needs a fix.
Example
Scene 1 welcome (30s). Scene 2 demo (45s). Scene 3 recap (20s). Each scene gets its own slide and avatar block.
Product names, acronyms, and uncommon terms may need phonetic hints or simplified phrasing.
Example
Script line: 'Open Northwind CRM' — note pronunciation: 'NORTH-wind C-R-M'. Avoid 'NwCRM' in speech.
Put lists, steps, and labels on screen; let the avatar introduce and transition, not read dense bullets verbatim.
Example
Avatar: 'Here are three steps.' On-screen numbered list appears while the avatar summarizes each step in one sentence.
Multiple languages and dubbing share your credit pool—finalize the primary script before multiplying locales.
Example
Lock English script v3, generate once, then dub to es-ES and fr-FR only. Do not dub draft v2.
Script in scenes, not one block
Break training and explainer content into scenes with clear durations. Scene boundaries are your best unit for fixes and credit control.
Optimize for spoken delivery
Use short sentences, active verbs, and natural transitions. If it sounds stiff when read aloud, the avatar will sound stiff too.
Lock compliance copy verbatim
Legal, safety, and policy lines should be pasted exactly as approved. Mark them in the script so they are not paraphrased during edits.
Choose avatar and voice to match audience
Stock avatars cover many presenter styles; personal avatars and voice clones are plan-gated. Pick one primary presenter per video for continuity.
Support speech with visuals
Show steps, diagrams, and labels on screen. Presenters work best as guides, not as readers of dense paragraphs.
Finalize before dubbing
Synthesia supports many languages and voices; AI dubbing draws from shared credits. Finish the primary-language script and timing before generating alternate languages.
Watch minutes and credits
Presenter videos are plan-capped by minutes and credits. Estimate runtime early, especially for batches and dubs.
Test clones and personal avatars small
Generate a short sample with voice cloning and personal avatars before committing to long-form content. Adjust pacing and sentence length based on that sample.
Use templates for repeatable formats
Monthly updates, personalized welcomes, and regional variants benefit from fixed layouts with controlled variable fields.
Integrate via API deliberately
Creator and Enterprise include API access. Version templates, validate required variables, and handle render failures before connecting HR, LMS, or marketing systems.
Pricing, features, limits and latest updates
AI presenter videos from a script, built for training and explainers.
Free / Paid · Free
Basic
Free
Starter
$29/ Monthly
Creator
$89/ Monthly
Enterprise
Custom
Creates presenter videos from a script; minutes and credits are plan-capped.
Includes 160+ languages and voices plus AI dubbing from the shared credit pool.
API access is included on Creator and Enterprise plans.
Voice cloning can be paired with a custom personal avatar.
Stock and personal AI avatars present the scripted video; counts vary by plan.
Share a prompt that worked for you. Username and email are shown with your submission. External links are not allowed.
Example prompt
Prompt
Assistant prompt: Create a 90-second training video explaining two-factor authentication for new employees. Tone: friendly and practical. Include one on-screen checklist scene and a plain-language recap.
Short explanation
Synthesia Assistant helps draft outline and script; final video generation consumes credits separately.
Example prompt
Prompt
Scene 1 (15s): Welcome and goal. Scene 2 (45s): Step-by-step setup with on-screen labels. Scene 3 (20s): Common mistakes and recap. Voice: clear presenter, neutral accent. No jargon without definition.
Short explanation
Script-first workflows work best with short scenes and visuals supporting speech, not dense paragraphs.
Required fields are marked. Variation, result, and explanation are optional.
AI video platform for text-to-video and image-to-video generation with credit-based plans and tools including Pikascenes, Pikadditions, Pikaswaps, and Pikatwists.
VideoFree / Paid
AI audio recording and editing on the web, including Enhance Speech for noise removal and Studio for browser-based podcast production.
Video