Multilo can now turn your project sources — research papers, PDFs, Word documents, notes, slides — into a fully narrated, animated video overview. No external service, no upload, no per-minute fee. Pick a format and a voice, and the agent reads your sources, drafts a script, narrates it on-device, and assembles a 1920×1080 video that opens in an in-app tab.
Three formats
- Cinematic — filmic look with letterbox bars, moody grading, kinetic typography, slow dissolves, film grain, and stronger Ken-Burns moves. Great for thesis defenses, lab open-day reels, and conference promo cuts.
- Explainer — clean, structured deck that connects the ideas across multiple sources. Good default for a paper walkthrough or a literature-review summary.
- Brief — a punchy, fast overview. A 60–90 second "what's in this paper" for a group chat, a lab Slack, or a supervisor preview.
What it ships
- Source-driven script — the agent reads every selected source, extracts the actual claims and figures, and writes a scene-by-scene script grounded in your library. No fabricated content, no "we asked ChatGPT what's in your paper" hallucinations.
- Figure intelligence — native chart re-rendering, panel-by-panel reveals on multi-panel figures, and draw-on animations for vector diagrams. Your real figures become real animated visuals, not screenshots.
- On-device narration — pick from seven curated voices (Warm, Bella, Nicole, Michael, Puck, Emma, George — US and UK, female and male). Powered by Kokoro running locally on your machine; no audio leaves the device.
- Arrow callouts and overlays — the script writer plans where to point your attention frame by frame; the renderer draws arrows and highlights at the right beats.
- Multilo watermark — light brand mark in the corner so any reshares stay attributable.
- Plays in an in-app tab — the finished video opens in a Multilo tab with native controls. Export to MP4 to share, present, or upload.
Steer the focus
Free-text focus field plus one-click chips to set the angle:
- Summarize the key findings
- Explain the methodology
- Highlight limitations & future work
- Make it beginner-friendly
Drop in your own focus and the script writer follows it.
How it works
- Open the kickoff dialog — pick Video Overview from the chat agents picker.
- Pick sources — any combination of PDFs, .docx, .pptx, slides, markdown, or plain text from your project.
- Choose format and voice — Cinematic / Explainer / Brief, plus one of seven on-device voices.
- Add focus (optional) — type a sentence or click a suggestion chip to steer what the script emphasizes.
- Watch it build — script → figure extraction → narration → render — every step shows live in the chat thread.
- Play in-tab and export — the finished video opens in a Multilo tab. Export to MP4 when you're ready to share.
Why it matters
A narrated video overview is the highest-bandwidth way to share what your paper is actually about. Reviewers, supervisors, and lab-mates don't always have 25 minutes for a full read — but they have two for a Cinematic teaser. Until now, making one meant Premiere, a voice actor, and a couple of evenings. Multilo cuts that to a single dialog, a couple of minutes of local render time, and a video that's actually grounded in your sources instead of hand-waved.
Use it for
- A 90-second teaser of your latest paper for your group chat
- A thesis-defense overview reel
- A literature-review video walking through five papers in one cut
- A conference promo for an accepted submission
- A study summary that lives in your project's
videos/folder
Try it
Open Multilo, pick a project with at least one paper or PDF, and choose Video Overview from the chat agents picker.