A collection of simple, standalone Python scripts for working with video, audio, image, and text data — designed for developers exploring multimodal AI.
Each utility lives in its own folder with examples and a CLI-friendly interface.
| Tool | Description |
|---|---|
segment_transcript_by_topic/ |
Transcribe and cluster audio/video by topic |
split_video_by_second/ |
Split a video file into N-second chunks |
extract_thumbnails/ |
Grab frames from a video every N seconds |
transcribe_audio/ |
Transcribe audio using Whisper |
search_local_media/ |
CLIP-based text search across your media folder |
scene_change_split/ |
Splits a video into separate clips based on detected scene changes. |
caption_search/ |
Searches for text within video captions or existing Whisper transcripts. |
generate_video_captions/ |
Generates SRT/VTT caption files from video or audio using Whisper. |
video_shot_segmenter/ |
Segments a video into individual shots based on visual changes. |
summarize_transcript/ |
Summarizes text from a transcript file (e.g., from Whisper output). |
blur_faces/ |
Detects and blurs faces in images or video frames for privacy. |
# Clone the repo
git clone https://github.com/mixpeek/multimodal-tools.git
cd multimodal-tools
# Pick a tool and follow its README
cd segment_transcript_by_topic
pip install -r requirements.txt
python segment_transcript.py --input path/to/video.mp4Multimodal content is everywhere — but tooling is scattered. This repo brings together focused, no-dependency-heavy scripts to help you get things done without setting up complex pipelines.
Ideal for:
- Prototyping and experimentation
- Content analysis workflows
- ML/AI feature extraction
- Exploring retrieval use cases
If you want to scale beyond local scripts, Mixpeek offers managed, production-ready multimodal extractors (video, image, audio, and more) you can plug into your stack.
Want to add a new tool or improve an existing one? PRs welcome.