Video Embeddings for Multimodal Search

embed

Vectors for video, built for production.

Marengo produces 512-dimensional embeddings across visual, audio, dialogue, and on-screen text in a single vector. Ready for semantic search, recommendations, RAG, and anomaly detection.

Try on Playground

From video to vectors, in one call.

Generate contextual vectors across every modality: visual, audio, spoken word, and on-screen text. A single embedding to power semantic search, recommendations, anomaly detection, and RAG pipelines.

Multimodal shouldn't mean multi-model.

One model handles image, audio, text, and video. No stitching, no separate pipelines, no orchestration between vendors for cross-modal queries.

Domain specific.

Marengo understands your domain vocabulary and customer-specific terminology. Embeddings reflect how your team and your buyers actually describe the world.

Faster processing. Better results.

Native video support reduces processing time, increases throughput and lowers cost. At 180x run time indexing, process 10,000 hours of video in less than an hour.

From video to vectors, in one call.

Generate contextual vectors across every modality: visual, audio, spoken word, and on-screen text. A single embedding to power semantic search, recommendations, anomaly detection, and RAG pipelines.

Multimodal shouldn't mean multi-model.

One model handles image, audio, text, and video. No stitching, no separate pipelines, no orchestration between vendors for cross-modal queries.

Domain specific.

Marengo understands your domain vocabulary and customer-specific terminology. Embeddings reflect how your team and your buyers actually describe the world.

Faster processing. Better results.

Native video support reduces processing time, increases throughput and lowers cost. At 180x run time indexing, process 10,000 hours of video in less than an hour.

Build with video embeddings

RAG pairing

Pair our models with your RAG pipeline to retrieve relevant information and improve data output.

High-quality training data

Transform workflows with embeddings to create training data, improve data quality, and reduce manual labeling needs.

Training models

Use embeddings to improve data quality when training large language models.

Anomaly detection

Identify anomalies – for example, detect and remove corrupt videos that only display a black background – to enhance data quality.

Build with video embeddings

RAG pairing

Pair our models with your RAG pipeline to retrieve relevant information and improve data output.

High-quality training data

Transform workflows with embeddings to create training data, improve data quality, and reduce manual labeling needs.

Training models

Use embeddings to improve data quality when training large language models.

Anomaly detection

Identify anomalies – for example, detect and remove corrupt videos that only display a black background – to enhance data quality.

Build with video embeddings

RAG pairing

Pair our models with your RAG pipeline to retrieve relevant information and improve data output.

High-quality training data

Transform workflows with embeddings to create training data, improve data quality, and reduce manual labeling needs.

Training models

Use embeddings to improve data quality when training large language models.

Anomaly detection

Identify anomalies – for example, detect and remove corrupt videos that only display a black background – to enhance data quality.

Build with video embeddings

RAG pairing

Pair our models with your RAG pipeline to retrieve relevant information and improve data output.

High-quality training data

Transform workflows with embeddings to create training data, improve data quality, and reduce manual labeling needs.

Training models

Use embeddings to improve data quality when training large language models.

Anomaly detection

Identify anomalies – for example, detect and remove corrupt videos that only display a black background – to enhance data quality.

Python

Node.js

import requests
 
# Step 2: Define the API URL and the specific endpoint
API_URL = "https://api.twelvelabs.io/v1.3"
INDEXES_URL = f"{API_URL}/indexes"
 
# Step 3: Create the necessary headers for authentication
headers = {
    "x-api-key": "<YOUR_API_KEY>"
}
 
# Step 4: Prepare the data payload for your API request
INDEX_NAME = "<YOUR_INDEX_NAME>"
data = {
    "models": [
        {
            "model_name": "marengo3.0",
            "model_options": ["visual", "audio"]
        }
    ]
}