TeX2Vid turns a LaTeX paper into a narrated, source-grounded video in minutes.
TeX2Vid converts a LaTeX source into a narrated video presentation. Every figure, result, claim, and attribution is grounded in the source paper — each spoken and on-screen claim traces back to the LaTeX source, so the video says what the source paper says.
| Format | Length | Aspect | Best for |
|---|---|---|---|
| Research Teaser | 30 s | 9:16 | Vertical hook for Shorts, Reels & TikTok |
| Research Short | 50 s | 9:16 | Setting up the problem and the key result |
| Research Highlight | 1 min | 16:9 | Horizontal walk-through on Beamer slides |
| Video Abstract | under 3 min | 16:9 | Full abstract for conferences and journals |
- 🔗 Grounded & verifiable — claims traced to source paper.
- 🎙️ AI narration or your own voice — record yourself or use AI-synthesized speech.
- 🎨 Configurable Beamer themes and researcher / general-public audience settings.
- 📚 Broad field support — CS, ML, NLP, CV, math, physics, statistics, EE, biology, economics, and more.
- 📄 Template-compatible — works with Nature, Science, PNAS, NeurIPS, ICML, ICLR, CVPR, ACL, IEEE, ACM, Springer, Elsevier, and others.
- Go to tex2vid.sheqai.com
- Upload your LaTeX source.
- Pick a video format (Research Teaser, Short, Highlight, or Video Abstract), audience, and voice (default AI voice or record your own voice).
- Render — then share to social media, conferences, or journals.
- You may consider submitting to ScholarReels; the video may be included if it meets the criteria set out in the Terms.