Searchable Video Library
Turn an unstructured video archive into a fully searchable library. Each video is decomposed into scenes with transcriptions, visual embeddings, and metadata. Users search by natural language and jump directly to the relevant moment in any video.
from mixpeek import Mixpeekclient = Mixpeek(api_key="YOUR_API_KEY")# Create video library collectioncollection = client.collections.create(namespace_id="ns_your_namespace",name="video_library",extractors=["multimodal-extractor", "text-extractor"],params={"video_chunking": "scene-based"})# Upload your video archiveclient.buckets.upload(bucket_id="bkt_videos",url="s3://your-bucket/video-archive/")# Build a search retrieverretriever = client.retrievers.create(namespace_id="ns_your_namespace",name="video_library_search",collection_ids=["col_video_library"],stages=[{"type": "feature_search", "top_k": 50},{"type": "rerank", "top_k": 10}])# Search for specific momentsresults = client.retrievers.execute(retriever_id=retriever["retriever_id"],query={"text": "product roadmap presentation Q3 goals"})for doc in results["results"]:print(f"Video: {doc['root_object_id']}")print(f" Timestamp: {doc['start_time']:.1f}s - {doc['end_time']:.1f}s")print(f" Transcript: {doc['text'][:100]}...")
Feature Extractors
Retriever Stages
rerank
Rerank documents using cross-encoder models for accurate relevance
