MathLens

To generate narrated math teaching videos, you can input a math problem as text or an image to trigger an automated pipeline. The workflow produces HTML visuals, storyboard scripts, TTS audio, and Manim animations for a complete teaching video output.

Yes, Manim animations can be synchronized with TTS audio. The workflow uses edge-tts to generate narration and applies per-scene timing data to ensure the visual Manim animations align directly with the speech audio.

The best way to turn a geometry problem image into a video is through an end-to-end pipeline that analyzes the image, generates HTML/SVG annotated diagrams for reasoning steps, and renders a final Manim animation with narration.

Yes, you need edge-tts and mutagen installed. Edge-tts generates the narration audio for the storyboard, while mutagen handles audio metadata to achieve per-scene audio-visual synchronization and timing validation.

Storyboard generation works by analyzing the math problem, creating an HTML visual explanation, and producing a timed narration script. This script scaffolds the Manim code generation and ensures TTS audio aligns with the visual scenes.

Limitations include dependency on edge-tts for audio generation and Manim for rendering, requiring a reproducible workflow with templates and scaffolding to validate timing and ensure outputs align correctly across different math problems.