mrspideynihal/lecture-video-maker

Lecture Video Maker is a Python tool that turns any text topic into a complete lecture video, including narration, subtitles, and stock footage. You type in a subject, and the tool generates a script using a local AI model, reads it aloud using a Microsoft speech engine, pairs it with relevant video clips or images from Pexels, and combines everything into a single MP4 file. The tool works without paying for AI services. The script writing uses Ollama, which runs AI models locally on your own machine. The voiceover comes from Microsoft Edge-TTS, which is also free. The only account required is a free Pexels API key to pull in stock video and image footage. Video assembly is handled by FFmpeg. You can control the tool two ways: through a web browser interface that opens at localhost:8080, or from the command line. The browser interface lets you set your Pexels key, pick an AI model, choose a voice accent, enter your topic, and watch the video being built in real time. Before the final render, you can review and edit the script segment by segment. The CLI option offers the same process through text prompts instead. Setup requires Python 3.8 or newer, FFmpeg installed on your system and available in your terminal, Ollama running with at least one language model pulled, and a free Pexels account. Once those are in place, install four Python packages and launch the script. The tool saves your settings between runs so you do not need to re-enter the Pexels key every time. Output lands in a folder called lecture_output, organized by topic and timestamp. Each run saves the script, individual audio clips, downloaded footage, subtitle frames, and the finished video. If Pexels returns no results for a topic, the tool falls back to a plain colored background rather than failing.