Your real UI, rebuilt in code
Claude reads your components and recreates the screens as animated React. No screen recordings, no blurry zooms, no real customer data on screen.
- Sharp at any resolution
- Clean sample data
- Change it without re-recording
Features
Product videos, built from the source.
DemoVexa connects the four jobs behind every product video (story, voice, animation and publishing) and hands each to the tool that does it best.
How it works
Five steps from repository
to a finished video
You stay in charge of the story. DemoVexa does the scripting, narration, animation and rendering, and shows its work at every step.
- 01
Connect your product
Point a workspace at your code folder, in any language, plus docs and modules. DemoVexa learns your screens, names and flows.
Code in any language
- 02
Write the brief
Say who the video is for and what it should show, or let Claude draft the brief from your code and a one-line idea.
Or start from an idea
- 03
Script and voice
Claude writes a scene-by-scene script. ElevenLabs narrates it with a native-sounding voice in any of 31 languages.
Word-level timing
- 04
Claude builds the video
A build agent recreates your UI in React and animates it with Remotion, every move timed to the words being spoken.
Your brand kit applied
- 05
Render and publish
Export 16:9, 9:16 and 1:1 MP4s with covers and ready-to-post copy for LinkedIn, Instagram, YouTube and your blog.
1080p MP4 · covers · copy
A brand kit per workspace
Logo, colours, fonts and the call to action, applied to every video of that product.
Batch builds, overnight
Queue a release's worth of videos before you leave. Wake up to rendered MP4s.
Never stare at a blank brief
The ideas generator reads your product and suggests videos worth making: launches, how-tos, comparisons.
Covers and publish copy
A cover per format plus titles, descriptions and hashtags written for each platform.
Projects that travel
Export any project as a single .demovexa file and import it on another machine, or hand it to a client.
Edit by asking
Change pace, wording, colours or the ending in plain words. Claude edits the video's code and you preview the result in seconds.
Event recaps from your photos
A second template for launches, meetups and conferences: drop in photos and clips, get a branded recap.
In detail
Everything in the box
Every plan includes every feature below; plans differ only in volume, seats and devices.
Workspaces
- One workspace per product: code folder in any language, docs and modules
- Brand kit: logo, colours, fonts and call to action
- Default language and voice per workspace
- Switch products from the sidebar or the command palette
Script and voice
- Claude drafts the brief from an idea, or you write it
- Scene-by-scene scripts with on-screen text and timing
- ElevenLabs voices with a voice library preview
- Word-level timing so animation lands on the spoken word
The build agent
- Recreates your screens as React components from your code
- Animates them with Remotion: camera moves, cursors, typing, transitions
- Uses clean sample data instead of real customer records
- Revise any scene by asking in plain words
Render and publish
- 1080p MP4 in 16:9, 9:16 and 1:1
- Covers for each format
- Posts, titles and hashtags for LinkedIn, Instagram and YouTube, plus a full blog post
- Preview renders while you iterate, full quality when you're done
Workflow
- Batch builds: queue many videos and let them run overnight
- Ideas generator for launches, how-tos and tours
- Export and import projects as .demovexa files
- Event-recap template from your photos and clips
Privacy and control
- Runs on your computer: macOS, Windows and Linux
- API keys stored in the OS keychain
- Bring your own Anthropic and ElevenLabs accounts
- Use a Claude API key or sign in with your Claude account
Multilingual
One product. 31 languages. Voices that sound local.
Claude writes each script natively, not word for word, and ElevenLabs reads it with a native-sounding voice. Product and UI names stay exactly as they appear in your app.
- Scenes re-time themselves to each voiceover
- Captions, title cards and the CTA translated too
- Batch one video into many languages overnight
Scene 5 · same line, five voices
ElevenLabs- EN
Send the invoice the moment the work is done.
0:03.4 - ES
Envía la factura en cuanto termina el trabajo.
0:03.7 - HI
काम पूरा होते ही इनवॉइस भेजें।
0:03.2 - JA
作業が終わったら、すぐに請求書を送信。
0:03.5 - AR
أرسل الفاتورة فور انتهاء العمل.
0:03.3
All 31 voiceover languages
- enEnglish
- hiHindi
- esSpanish
- frFrench
- deGerman
- ptPortuguese
- itItalian
- nlDutch
- plPolish
- svSwedish
- daDanish
- noNorwegian
- fiFinnish
- trTurkish
- ruRussian
- ukUkrainian
- arArabic
- jaJapanese
- koKorean
- zhChinese (Mandarin)
- idIndonesian
- msMalay
- viVietnamese
- filFilipino
- taTamil
- bnBengali
- urUrdu
- elGreek
- csCzech
- roRomanian
- huHungarian
Formats
One build. Three formats, each composed for its frame.
Vertical and square aren't crops of the wide video. Each layout is arranged for its own screen, with captions placed where they read best.
16:91920 × 1080
YouTube, your site, LinkedIn
9:161080 × 1920
TikTok, Reels, Shorts
1:11080 × 1080
Instagram, LinkedIn and X feeds
Local-first
Your code stays on your computer.
DemoVexa is a desktop app, not a cloud service that ingests your repository. The parts a video needs go straight from your machine to the AI you connect.
Runs on your computer
Your code, voiceovers, projects and renders live on your disk. DemoVexa has no servers that store them.
Keys in your OS keychain
Anthropic and ElevenLabs keys sit in the macOS Keychain, Windows Credential Manager or Secret Service.
Bring your own keys
Claude and ElevenLabs bill you directly at their prices. No mark-up, no credits to buy from us.
Nothing goes to DemoVexa
Only what a build needs is sent, straight to Claude through your own account. DemoVexa never sees it.
Your next release deserves a video.
Point DemoVexa at your code and have a narrated, on-brand product video before the end of the day. 14 days free.