Two avatars, one conversation → they take turns speaking → a finished side-by-side video, in your voice and theirs
Create a free account or sign in to generate split-screen conversation videos.
Drag & drop a video file here, or click to browse
MP4, MOV, AVI, FLV accepted · Max 500MB
Works for YouTube videos (needs captions/subtitles enabled) and most articles/blogs. Fetched text lands in the box below — review it before processing.
Add a transcript, a prompt, or both. With a prompt only, the AI writes the script from scratch.
⚠️ Native Bhojpuri/Maithili voices aren't available — the script will be written in that language, but spoken narration uses the closest available voice (usually Hindi), so pronunciation won't be perfect.
Uncheck this if you just want the cleaned script/transcript first. Video generation costs about $1 per minute of finished video.
Works for every source: an uploaded video, a transcript, a fetched link, an edited transcript, or a cinematic clip. Pick what sits behind the avatar below.
Tip: a slide image works well for the "lecture on screen, presenter in the corner" look.
The premium avatar looks best but costs most. For a small corner face, the budget or voice+photo options look fine at a fraction of the price.
The avatar narrates the whole script and nothing is cut — this controls how long the corner box stays visible.
Uses a clear, front-facing photo — it becomes a talking avatar you can pick above. Saved to your account, so it stays after a refresh and shows up on any browser you sign in from. For a full live-face video clone instead of a photo, a one-time consent recording is required on the provider's site — do that there first, then paste the avatar ID below.
There are different ID types depending on how you made it — a Digital Twin (video-clone) look is an Avatar ID; a Photo Avatar Group look (the "Choose Avatar & Look" screen with "Upload look"/"Design with AI") is a Talking Photo ID. Pick the matching type below — if generation fails, try the other one.
Sending to AI...
All personal information has been removed. Your clean script:
Edit the transcript here before generating — the saved version is what gets narrated, and the cost estimate below updates to match.
What do you want to make with this script?
Pick an AI presenter and voice above, then generate — or it already started automatically if the checkbox above was ticked.
This is a preview of the payment gate that will require acceptance + payment here once Razorpay is connected. Generation below still works directly for now.
Submitting…
Watch it back — does this look right?
Free plan: 5 script cleanings/month.
Upgrade for unlimited →
Turn this script into YouTube Shorts or a full long-form upload, with AI-generated music and other tools.
🔧 Coming soonDifferent from the avatar tool above — no script, no face. Describe a scene and the AI renders an actual cinematic clip from scratch: settings, camera movement, action.
Built-in audio adds roughly a third to the cost. If your avatar is narrating (or you'll add sound later), pick the cheaper "No added audio" option — the estimate below updates to match.
Uses the avatar, voice and saved script from the section above. The cinematic clip renders first, then your avatar (Digital Twin or Photo Avatar) is composited into the corner you pick — this adds the avatar cost on top of the cinematic cost.
This is a preview of the payment gate that will require acceptance + payment here once Razorpay is connected. Generation below still works directly for now.
Costs are shown for your approval first — nothing is generated until you confirm.
Submitting…
Watch it back — does this look right?
Build a longer film from multiple scenes and characters. Each character is one of your saved avatars with a chosen voice and language; name them in your script and assign each a scene. Scenes render one by one and stitch into a full film — so you can go well past the short-clip limit, up to a 2-minute+ short.
Give each a NAME (use the same name in your script), pick a saved avatar, a voice, and a language.
No saved avatars yet? Create or save one in the “My Avatars” section above first — they'll appear here automatically.
Format: NAME: their dialogue on each line. Wrap a scene/background description in square brackets on its own line, e.g. [a rainy neon street at night], for a cinematic cutaway. Then click Parse to turn it into editable scenes.
No scenes yet — parse a screenplay above, or add a blank scene.
0 scenes · ~0 sec of film
Tip: keep scenes on Fast/draft + silent while you're finding the story, then raise quality on the final render. Rendering happens scene-by-scene, so a long film simply takes longer — you'll see progress below.
Starting…
Paste your sources into a research/notes tool to get summaries/podcasts, then bring the output back here as your transcript or prompt.
Open research tool →Two avatars side by side — the AI on the left, you on the right, both on screen the whole time.
The avatar lip-syncs to this recording, so the AI text box below is ignored in this mode.
Label each line with AI: or ME: — they speak one after the other, in this order. Turn mode uses the synthetic AI voice.
3 credits = ₹100 (incl. taxes) ≈ about 10 seconds of video. This is an estimate; final credits are based on the actual video length.
Starting…