Mixpeek Logo
    Enrichment

    Multimodal Enrichment Pipeline

    Multi-tier collections that extract vision, audio, text, and metadata signals. This is the backbone—most enterprise pipelines start here.

    video
    image
    audio
    text
    Multi-Tier
    142.0K runs
    Deploy Recipe

    Why This Matters

    Enrichment pipelines are infrastructure. Raw → embeddings → searchable documents. Define once, query forever.

    from mixpeek import Mixpeek
    client = Mixpeek(api_key="your-api-key")
    # Create multi-tier enrichment collection
    collection = client.collections.create(
    collection_name="enriched_media",
    feature_extractor={
    "feature_extractor_name": "multimodal_extractor",
    "version": "v1",
    "parameters": {
    "enable_transcription": True,
    "enable_object_detection": True
    }
    }
    )
    # Index objects (triggers extraction pipeline)
    client.buckets.objects.create(
    bucket_id="raw-media",
    blobs=[{
    "property": "video",
    "url": "s3://bucket/meeting-recording.mp4"
    }]
    )
    # Search enriched content
    results = client.retrievers.execute(
    retriever_id="enriched-search",
    inputs={
    "query_text": "quarterly roadmap discussion",
    "filters": {
    "metadata.speaker": "CEO"
    }
    }
    )

    Retrieval Flow

    1

    Search enriched features

    2

    Filter by extracted metadata

    Feature Extractors

    Image Embedding

    Generate visual embeddings for similarity search and clustering

    752K runs

    Video Embedding

    Generate vector embeddings for video content

    610K runs

    Audio Transcription

    Transcribe audio content to text

    450K runs

    Text Embedding

    Extract semantic embeddings from documents, transcripts and text content

    827K runs

    Object Detection

    Identify and locate objects within images with bounding boxes

    631K runs

    Retriever Stages

    feature search

    Search collections using multimodal embeddings

    search

    attribute filter

    Filter documents by metadata attributes

    filter

    Documentation