This post was originally published on this site

Imagine building AI applications that deliver accurate, current information without the complexity of developing intricate data retrieval systems. Today, we’re excited to announce the general availability of Web Grounding, a new built-in tool for Nova models on Amazon Bedrock.

Web Grounding provides developers with a turnkey Retrieval Augmented Generation (RAG) option that allows the Amazon Nova foundation models to intelligently decide when to retrieve and incorporate relevant up-to-date information based on the context of the prompt. This helps to ground the model output by incorporating cited public sources as context, aiming to reduce hallucinations and improve accuracy.

When should developers use Web Grounding?

Developers should consider using Web Grounding when building applications that require access to current, factual information or need to provide well-cited responses. The capability is particularly valuable across a range of applications, from knowledge-based chat assistants providing up-to-date information about products and services, to content generation tools requiring fact-checking and source verification. It’s also ideal for research assistants that need to synthesize information from multiple current sources, as well as customer support applications where accuracy and verifiability are crucial.

Web Grounding is especially useful when you need to reduce hallucinations in your AI applications or when your use case requires transparent source attribution. Because it automatically handles the retrieval and integration of information, it’s an efficient solution for developers who want to focus on building their applications rather than managing complex RAG implementations.

Getting started
Web Grounding seamlessly integrates with supported Amazon Nova models to handle information retrieval and processing during inference. This eliminates the need to build and maintain complex RAG pipelines, while also providing source attributions that verify the origin of information.

Let’s see an example of asking a question to Nova Premier using Python to call the Amazon Bedrock Converse API with Web Grounding enabled.

First, I created an Amazon Bedrock client using AWS SDK for Python (Boto3) in the usual way. For good practice, I’m using a session, which helps to group configurations and make them reusable. I then create a BedrockRuntimeClient.

try:
    session = boto3.Session(region_name='us-east-1')
    client = session.client(
        'bedrock-runtime')

I then prepare the Amazon Bedrock Converse API payload. It includes a “role” parameter set to “user”, indicating that the message comes from our application’s user (compared to “assistant” for AI-generated responses).

For this demo, I chose the question “What are the current AWS Regions and their locations?” This was selected intentionally because it requires current information, making it useful to demonstrate how Amazon Nova can automatically invoke searches using Web Grounding when it determines that up-to-date knowledge is needed.

# Prepare the conversation in the format expected by Bedrock
question = "What are the current AWS regions and their locations?"
conversation = [
   {
     "role": "user",  # Indicates this message is from the user
     "content": [{"text": question}],  # The actual question text
      }
    ]

First, let’s see what the output is without Web Grounding. I make a call to Amazon Bedrock Converse API.

# Make the API call to Bedrock 
model_id = "us.amazon.nova-premier-v1:0" 
response = client.converse( 
    modelId=model_id, # Which AI model to use 
    messages=conversation, # The conversation history (just our question in this case) 
    )
print(response['output']['message']['content'][0]['text'])

I get a list of all the current AWS Regions and their locations.

Now let’s use Web Grounding. I make a similar call to the Amazon Bedrock Converse API, but declare nova_grounding as one of the tools available to the model.

model_id = "us.amazon.nova-premier-v1:0" 
response = client.converse( 
    modelId=model_id, 
    messages=conversation, 
    toolConfig= {
          "tools":[ 
              {
                "systemTool": {
                   "name": "nova_grounding" # Enables the model to search real-time information
                 }
              }
          ]
     }
)

After processing the response, I can see that the model used Web Grounding to access up-to-date information. The output includes reasoning traces that I can use to follow its thought process and see where it automatically queried external sources. The content of the responses from these external calls appear as [HIDDEN] – a standard practice in AI systems that both protects sensitive information and helps manage output size.

Additionally, the output also includes citationsContent objects containing information about the sources queried by Web Grounding.

Finally, I can see the list of AWS Regions. It finishes with a message right at the end stating that “These are the most current and active AWS regions globally.”

Web Grounding represents a significant step forward in making AI applications more reliable and current with minimum effort. Whether you’re building customer service chat assistants that need to provide up-to-date accurate information, developing research applications that analyze and synthesize information from multiple sources, or creating travel applications that deliver the latest details about destinations and accommodations, Web Grounding can help you deliver more accurate and relevant responses to your users with a convenient turnkey solution that is straightforward to configure and use.

Things to know
Amazon Nova Web Grounding is available today in US East (N. Virginia). Web Grounding will also soon launch on US East (Ohio), and US West (Oregon).

Web Grounding incurs additional cost. Refer to the Amazon Bedrock pricing page for more details.

Currently, you can only use Web Grounding with Nova Premier but support for other Nova models will be added soon.

If you haven’t used Amazon Nova before or are looking to go deeper, try this self-paced online workshop where you can learn how to effectively use Amazon Nova foundation models and related features for text, image, and video processing through hands-on exercises.

Matheus Guimaraes | @codingmatheus

This post was originally published on this site

Today, we’re introducing Amazon Nova Multimodal Embeddings, a state-of-the-art multimodal embedding model for agentic retrieval-augmented generation (RAG) and semantic search applications, available in Amazon Bedrock. It is the first unified embedding model that supports text, documents, images, video, and audio through a single model to enable crossmodal retrieval with leading accuracy.

Embedding models convert textual, visual, and audio inputs into numerical representations called embeddings. These embeddings capture the meaning of the input in a way that AI systems can compare, search, and analyze, powering use cases such as semantic search and RAG.

Organizations are increasingly seeking solutions to unlock insights from the growing volume of unstructured data that is spread across text, image, document, video, and audio content. For example, an organization might have product images, brochures that contain infographics and text, and user-uploaded video clips. Embedding models are able to unlock value from unstructured data, however traditional models are typically specialized to handle one content type. This limitation drives customers to either build complex crossmodal embedding solutions or restrict themselves to use cases focused on a single content type. The problem also applies to mixed-modality content types such as documents with interleaved text and images or video with visual, audio, and textual elements where existing models struggle to capture crossmodal relationships effectively.

Nova Multimodal Embeddings supports a unified semantic space for text, documents, images, video, and audio for use cases such as crossmodal search across mixed-modality content, searching with a reference image, and retrieving visual documents.

Evaluating Amazon Nova Multimodal Embeddings performance
We evaluated the model on a broad range of benchmarks, and it delivers leading accuracy out-of-the-box as described in the following table.

Nova Multimodal Embeddings supports a context length of up to 8K tokens, text in up to 200 languages, and accepts inputs via synchronous and asynchronous APIs. Additionally, it supports segmentation (also known as “chunking”) to partition long-form text, video, or audio content into manageable segments, generating embeddings for each portion. Lastly, the model offers four output embedding dimensions, trained using Matryoshka Representation Learning (MRL) that enables low-latency end-to-end retrieval with minimal accuracy changes.

Let’s see how the new model can be used in practice.

Using Amazon Nova Multimodal Embeddings
Getting started with Nova Multimodal Embeddings follows the same pattern as other models in Amazon Bedrock. The model accepts text, documents, images, video, or audio as input and returns numerical embeddings that you can use for semantic search, similarity comparison, or RAG.

Here’s a practical example using the AWS SDK for Python (Boto3) that shows how to create embeddings from different content types and store them for later retrieval. For simplicity, I’ll use Amazon S3 Vectors, a cost-optimized storage with native support for storing and querying vectors at any scale, to store and search the embeddings.

Let’s start with the fundamentals: converting text into embeddings. This example shows how to transform a simple text description into a numerical representation that captures its semantic meaning. These embeddings can later be compared with embeddings from documents, images, videos, or audio to find related content.

To make the code easy to follow, I’ll show a section of the script at a time. The full script is included at the end of this walkthrough.

import json
import base64
import time
import boto3

MODEL_ID = "amazon.nova-2-multimodal-embeddings-v1:0"
EMBEDDING_DIMENSION = 3072

# Initialize Amazon Bedrock Runtime client
bedrock_runtime = boto3.client("bedrock-runtime", region_name="us-east-1")

print(f"Generating text embedding with {MODEL_ID} ...")

# Text to embed
text = "Amazon Nova is a multimodal foundation model"

# Create embedding
request_body = {
    "taskType": "SINGLE_EMBEDDING",
    "singleEmbeddingParams": {
        "embeddingPurpose": "GENERIC_INDEX",
        "embeddingDimension": EMBEDDING_DIMENSION,
        "text": {"truncationMode": "END", "value": text},
    },
}

response = bedrock_runtime.invoke_model(
    body=json.dumps(request_body),
    modelId=MODEL_ID,
    contentType="application/json",
)

# Extract embedding
response_body = json.loads(response["body"].read())
embedding = response_body["embeddings"][0]["embedding"]

print(f"Generated embedding with {len(embedding)} dimensions")

Now we’ll process visual content using the same embedding space using a photo.jpg file in the same folder as the script. This demonstrates the power of multimodality: Nova Multimodal Embeddings is able to capture both textual and visual context into a single embedding that provides enhanced understanding of the document.

Nova Multimodal Embeddings can generate embeddings that are optimized for how they are being used. When indexing for a search or retrieval use case, embeddingPurpose can be set to GENERIC_INDEX. For the query step, embeddingPurpose can be set depending on the type of item to be retrieved. For example, when retrieving documents, embeddingPurpose can be set to DOCUMENT_RETRIEVAL.

# Read and encode image
print(f"Generating image embedding with {MODEL_ID} ...")

with open("photo.jpg", "rb") as f:
    image_bytes = base64.b64encode(f.read()).decode("utf-8")

# Create embedding
request_body = {
    "taskType": "SINGLE_EMBEDDING",
    "singleEmbeddingParams": {
        "embeddingPurpose": "GENERIC_INDEX",
        "embeddingDimension": EMBEDDING_DIMENSION,
        "image": {
            "format": "jpeg",
            "source": {"bytes": image_bytes}
        },
    },
}

response = bedrock_runtime.invoke_model(
    body=json.dumps(request_body),
    modelId=MODEL_ID,
    contentType="application/json",
)

# Extract embedding
response_body = json.loads(response["body"].read())
embedding = response_body["embeddings"][0]["embedding"]

print(f"Generated embedding with {len(embedding)} dimensions")

To process video content, I use the asynchronous API. That’s a requirement for videos that are larger than 25MB when encoded as Base64. First, I upload a local video to an S3 bucket in the same AWS Region.

aws s3 cp presentation.mp4 s3://my-video-bucket/videos/

This example shows how to extract embeddings from both visual and audio components of a video file. The segmentation feature breaks longer videos into manageable chunks, making it practical to search through hours of content efficiently.

# Initialize Amazon S3 client
s3 = boto3.client("s3", region_name="us-east-1")

print(f"Generating video embedding with {MODEL_ID} ...")

# Amazon S3 URIs
S3_VIDEO_URI = "s3://my-video-bucket/videos/presentation.mp4"
S3_EMBEDDING_DESTINATION_URI = "s3://my-embedding-destination-bucket/embeddings-output/"

# Create async embedding job for video with audio
model_input = {
    "taskType": "SEGMENTED_EMBEDDING",
    "segmentedEmbeddingParams": {
        "embeddingPurpose": "GENERIC_INDEX",
        "embeddingDimension": EMBEDDING_DIMENSION,
        "video": {
            "format": "mp4",
            "embeddingMode": "AUDIO_VIDEO_COMBINED",
            "source": {
                "s3Location": {"uri": S3_VIDEO_URI}
            },
            "segmentationConfig": {
                "durationSeconds": 15  # Segment into 15-second chunks
            },
        },
    },
}

response = bedrock_runtime.start_async_invoke(
    modelId=MODEL_ID,
    modelInput=model_input,
    outputDataConfig={
        "s3OutputDataConfig": {
            "s3Uri": S3_EMBEDDING_DESTINATION_URI
        }
    },
)

invocation_arn = response["invocationArn"]
print(f"Async job started: {invocation_arn}")

# Poll until job completes
print("nPolling for job completion...")
while True:
    job = bedrock_runtime.get_async_invoke(invocationArn=invocation_arn)
    status = job["status"]
    print(f"Status: {status}")

    if status != "InProgress":
        break
    time.sleep(15)

# Check if job completed successfully
if status == "Completed":
    output_s3_uri = job["outputDataConfig"]["s3OutputDataConfig"]["s3Uri"]
    print(f"nSuccess! Embeddings at: {output_s3_uri}")

    # Parse S3 URI to get bucket and prefix
    s3_uri_parts = output_s3_uri[5:].split("/", 1)  # Remove "s3://" prefix
    bucket = s3_uri_parts[0]
    prefix = s3_uri_parts[1] if len(s3_uri_parts) > 1 else ""

    # AUDIO_VIDEO_COMBINED mode outputs to embedding-audio-video.jsonl
    # The output_s3_uri already includes the job ID, so just append the filename
    embeddings_key = f"{prefix}/embedding-audio-video.jsonl".lstrip("/")

    print(f"Reading embeddings from: s3://{bucket}/{embeddings_key}")

    # Read and parse JSONL file
    response = s3.get_object(Bucket=bucket, Key=embeddings_key)
    content = response['Body'].read().decode('utf-8')

    embeddings = []
    for line in content.strip().split('n'):
        if line:
            embeddings.append(json.loads(line))

    print(f"nFound {len(embeddings)} video segments:")
    for i, segment in enumerate(embeddings):
        print(f"  Segment {i}: {segment.get('startTime', 0):.1f}s - {segment.get('endTime', 0):.1f}s")
        print(f"    Embedding dimension: {len(segment.get('embedding', []))}")
else:
    print(f"nJob failed: {job.get('failureMessage', 'Unknown error')}")

With our embeddings generated, we need a place to store and search them efficiently. This example demonstrates setting up a vector store using Amazon S3 Vectors, which provides the infrastructure needed for similarity search at scale. Think of this as creating a searchable index where semantically similar content naturally clusters together. When adding an embedding to the index, I use the metadata to specify the original format and the content being indexed.

# Initialize Amazon S3 Vectors client
s3vectors = boto3.client("s3vectors", region_name="us-east-1")

# Configuration
VECTOR_BUCKET = "my-vector-store"
INDEX_NAME = "embeddings"

# Create vector bucket and index (if they don't exist)
try:
    s3vectors.get_vector_bucket(vectorBucketName=VECTOR_BUCKET)
    print(f"Vector bucket {VECTOR_BUCKET} already exists")
except s3vectors.exceptions.NotFoundException:
    s3vectors.create_vector_bucket(vectorBucketName=VECTOR_BUCKET)
    print(f"Created vector bucket: {VECTOR_BUCKET}")

try:
    s3vectors.get_index(vectorBucketName=VECTOR_BUCKET, indexName=INDEX_NAME)
    print(f"Vector index {INDEX_NAME} already exists")
except s3vectors.exceptions.NotFoundException:
    s3vectors.create_index(
        vectorBucketName=VECTOR_BUCKET,
        indexName=INDEX_NAME,
        dimension=EMBEDDING_DIMENSION,
        dataType="float32",
        distanceMetric="cosine"
    )
    print(f"Created index: {INDEX_NAME}")

texts = [
    "Machine learning on AWS",
    "Amazon Bedrock provides foundation models",
    "S3 Vectors enables semantic search"
]

print(f"nGenerating embeddings for {len(texts)} texts...")

# Generate embeddings using Amazon Nova for each text
vectors = []
for text in texts:
    response = bedrock_runtime.invoke_model(
        body=json.dumps({
            "taskType": "SINGLE_EMBEDDING",
            "singleEmbeddingParams": {
                "embeddingDimension": EMBEDDING_DIMENSION,
                "text": {"truncationMode": "END", "value": text}
            }
        }),
        modelId=MODEL_ID,
        accept="application/json",
        contentType="application/json"
    )

    response_body = json.loads(response["body"].read())
    embedding = response_body["embeddings"][0]["embedding"]

    vectors.append({
        "key": f"text:{text[:50]}",  # Unique identifier
        "data": {"float32": embedding},
        "metadata": {"type": "text", "content": text}
    })
    print(f"  ✓ Generated embedding for: {text}")

# Add all vectors to store in a single call
s3vectors.put_vectors(
    vectorBucketName=VECTOR_BUCKET,
    indexName=INDEX_NAME,
    vectors=vectors
)

print(f"nSuccessfully added {len(vectors)} vectors to the store in one put_vectors call!")

This final example demonstrates the capability of searching across different content types with a single query, finding the most similar content regardless of whether it originated from text, images, videos, or audio. The distance scores help you understand how closely related the results are to your original query.

# Text to query
query_text = "foundation models"  

print(f"nGenerating embeddings for query '{query_text}' ...")

# Generate embeddings
response = bedrock_runtime.invoke_model(
    body=json.dumps({
        "taskType": "SINGLE_EMBEDDING",
        "singleEmbeddingParams": {
            "embeddingPurpose": "GENERIC_RETRIEVAL",
            "embeddingDimension": EMBEDDING_DIMENSION,
            "text": {"truncationMode": "END", "value": query_text}
        }
    }),
    modelId=MODEL_ID,
    accept="application/json",
    contentType="application/json"
)

response_body = json.loads(response["body"].read())
query_embedding = response_body["embeddings"][0]["embedding"]

print(f"Searching for similar embeddings...n")

# Search for top 5 most similar vectors
response = s3vectors.query_vectors(
    vectorBucketName=VECTOR_BUCKET,
    indexName=INDEX_NAME,
    queryVector={"float32": query_embedding},
    topK=5,
    returnDistance=True,
    returnMetadata=True
)

# Display results
print(f"Found {len(response['vectors'])} results:n")
for i, result in enumerate(response["vectors"], 1):
    print(f"{i}. {result['key']}")
    print(f"   Distance: {result['distance']:.4f}")
    if result.get("metadata"):
        print(f"   Metadata: {result['metadata']}")
    print()

Crossmodal search is one of the key advantages of multimodal embeddings. With crossmodal search, you can query with text and find relevant images. You can also search for videos using text descriptions, find audio clips that match certain topics, or discover documents based on their visual and textual content. For your reference, the full script with all previous examples merged together is here:

import json
import base64
import time
import boto3

MODEL_ID = "amazon.nova-2-multimodal-embeddings-v1:0"
EMBEDDING_DIMENSION = 3072

# Initialize Amazon Bedrock Runtime client
bedrock_runtime = boto3.client("bedrock-runtime", region_name="us-east-1")

print(f"Generating text embedding with {MODEL_ID} ...")

# Text to embed
text = "Amazon Nova is a multimodal foundation model"

# Create embedding
request_body = {
    "taskType": "SINGLE_EMBEDDING",
    "singleEmbeddingParams": {
        "embeddingPurpose": "GENERIC_INDEX",
        "embeddingDimension": EMBEDDING_DIMENSION,
        "text": {"truncationMode": "END", "value": text},
    },
}

response = bedrock_runtime.invoke_model(
    body=json.dumps(request_body),
    modelId=MODEL_ID,
    contentType="application/json",
)

# Extract embedding
response_body = json.loads(response["body"].read())
embedding = response_body["embeddings"][0]["embedding"]

print(f"Generated embedding with {len(embedding)} dimensions")
# Read and encode image
print(f"Generating image embedding with {MODEL_ID} ...")

with open("photo.jpg", "rb") as f:
    image_bytes = base64.b64encode(f.read()).decode("utf-8")

# Create embedding
request_body = {
    "taskType": "SINGLE_EMBEDDING",
    "singleEmbeddingParams": {
        "embeddingPurpose": "GENERIC_INDEX",
        "embeddingDimension": EMBEDDING_DIMENSION,
        "image": {
            "format": "jpeg",
            "source": {"bytes": image_bytes}
        },
    },
}

response = bedrock_runtime.invoke_model(
    body=json.dumps(request_body),
    modelId=MODEL_ID,
    contentType="application/json",
)

# Extract embedding
response_body = json.loads(response["body"].read())
embedding = response_body["embeddings"][0]["embedding"]

print(f"Generated embedding with {len(embedding)} dimensions")
# Initialize Amazon S3 client
s3 = boto3.client("s3", region_name="us-east-1")

print(f"Generating video embedding with {MODEL_ID} ...")

# Amazon S3 URIs
S3_VIDEO_URI = "s3://my-video-bucket/videos/presentation.mp4"

# Amazon S3 output bucket and location
S3_EMBEDDING_DESTINATION_URI = "s3://my-video-bucket/embeddings-output/"

# Create async embedding job for video with audio
model_input = {
    "taskType": "SEGMENTED_EMBEDDING",
    "segmentedEmbeddingParams": {
        "embeddingPurpose": "GENERIC_INDEX",
        "embeddingDimension": EMBEDDING_DIMENSION,
        "video": {
            "format": "mp4",
            "embeddingMode": "AUDIO_VIDEO_COMBINED",
            "source": {
                "s3Location": {"uri": S3_VIDEO_URI}
            },
            "segmentationConfig": {
                "durationSeconds": 15  # Segment into 15-second chunks
            },
        },
    },
}

response = bedrock_runtime.start_async_invoke(
    modelId=MODEL_ID,
    modelInput=model_input,
    outputDataConfig={
        "s3OutputDataConfig": {
            "s3Uri": S3_EMBEDDING_DESTINATION_URI
        }
    },
)

invocation_arn = response["invocationArn"]
print(f"Async job started: {invocation_arn}")

# Poll until job completes
print("nPolling for job completion...")
while True:
    job = bedrock_runtime.get_async_invoke(invocationArn=invocation_arn)
    status = job["status"]
    print(f"Status: {status}")

    if status != "InProgress":
        break
    time.sleep(15)

# Check if job completed successfully
if status == "Completed":
    output_s3_uri = job["outputDataConfig"]["s3OutputDataConfig"]["s3Uri"]
    print(f"nSuccess! Embeddings at: {output_s3_uri}")

    # Parse S3 URI to get bucket and prefix
    s3_uri_parts = output_s3_uri[5:].split("/", 1)  # Remove "s3://" prefix
    bucket = s3_uri_parts[0]
    prefix = s3_uri_parts[1] if len(s3_uri_parts) > 1 else ""

    # AUDIO_VIDEO_COMBINED mode outputs to embedding-audio-video.jsonl
    # The output_s3_uri already includes the job ID, so just append the filename
    embeddings_key = f"{prefix}/embedding-audio-video.jsonl".lstrip("/")

    print(f"Reading embeddings from: s3://{bucket}/{embeddings_key}")

    # Read and parse JSONL file
    response = s3.get_object(Bucket=bucket, Key=embeddings_key)
    content = response['Body'].read().decode('utf-8')

    embeddings = []
    for line in content.strip().split('n'):
        if line:
            embeddings.append(json.loads(line))

    print(f"nFound {len(embeddings)} video segments:")
    for i, segment in enumerate(embeddings):
        print(f"  Segment {i}: {segment.get('startTime', 0):.1f}s - {segment.get('endTime', 0):.1f}s")
        print(f"    Embedding dimension: {len(segment.get('embedding', []))}")
else:
    print(f"nJob failed: {job.get('failureMessage', 'Unknown error')}")
# Initialize Amazon S3 Vectors client
s3vectors = boto3.client("s3vectors", region_name="us-east-1")

# Configuration
VECTOR_BUCKET = "my-vector-store"
INDEX_NAME = "embeddings"

# Create vector bucket and index (if they don't exist)
try:
    s3vectors.get_vector_bucket(vectorBucketName=VECTOR_BUCKET)
    print(f"Vector bucket {VECTOR_BUCKET} already exists")
except s3vectors.exceptions.NotFoundException:
    s3vectors.create_vector_bucket(vectorBucketName=VECTOR_BUCKET)
    print(f"Created vector bucket: {VECTOR_BUCKET}")

try:
    s3vectors.get_index(vectorBucketName=VECTOR_BUCKET, indexName=INDEX_NAME)
    print(f"Vector index {INDEX_NAME} already exists")
except s3vectors.exceptions.NotFoundException:
    s3vectors.create_index(
        vectorBucketName=VECTOR_BUCKET,
        indexName=INDEX_NAME,
        dimension=EMBEDDING_DIMENSION,
        dataType="float32",
        distanceMetric="cosine"
    )
    print(f"Created index: {INDEX_NAME}")

texts = [
    "Machine learning on AWS",
    "Amazon Bedrock provides foundation models",
    "S3 Vectors enables semantic search"
]

print(f"nGenerating embeddings for {len(texts)} texts...")

# Generate embeddings using Amazon Nova for each text
vectors = []
for text in texts:
    response = bedrock_runtime.invoke_model(
        body=json.dumps({
            "taskType": "SINGLE_EMBEDDING",
            "singleEmbeddingParams": {
                "embeddingPurpose": "GENERIC_INDEX",
                "embeddingDimension": EMBEDDING_DIMENSION,
                "text": {"truncationMode": "END", "value": text}
            }
        }),
        modelId=MODEL_ID,
        accept="application/json",
        contentType="application/json"
    )

    response_body = json.loads(response["body"].read())
    embedding = response_body["embeddings"][0]["embedding"]

    vectors.append({
        "key": f"text:{text[:50]}",  # Unique identifier
        "data": {"float32": embedding},
        "metadata": {"type": "text", "content": text}
    })
    print(f"  ✓ Generated embedding for: {text}")

# Add all vectors to store in a single call
s3vectors.put_vectors(
    vectorBucketName=VECTOR_BUCKET,
    indexName=INDEX_NAME,
    vectors=vectors
)

print(f"nSuccessfully added {len(vectors)} vectors to the store in one put_vectors call!")
# Text to query
query_text = "foundation models"  

print(f"nGenerating embeddings for query '{query_text}' ...")

# Generate embeddings
response = bedrock_runtime.invoke_model(
    body=json.dumps({
        "taskType": "SINGLE_EMBEDDING",
        "singleEmbeddingParams": {
            "embeddingPurpose": "GENERIC_RETRIEVAL",
            "embeddingDimension": EMBEDDING_DIMENSION,
            "text": {"truncationMode": "END", "value": query_text}
        }
    }),
    modelId=MODEL_ID,
    accept="application/json",
    contentType="application/json"
)

response_body = json.loads(response["body"].read())
query_embedding = response_body["embeddings"][0]["embedding"]

print(f"Searching for similar embeddings...n")

# Search for top 5 most similar vectors
response = s3vectors.query_vectors(
    vectorBucketName=VECTOR_BUCKET,
    indexName=INDEX_NAME,
    queryVector={"float32": query_embedding},
    topK=5,
    returnDistance=True,
    returnMetadata=True
)

# Display results
print(f"Found {len(response['vectors'])} results:n")
for i, result in enumerate(response["vectors"], 1):
    print(f"{i}. {result['key']}")
    print(f"   Distance: {result['distance']:.4f}")
    if result.get("metadata"):
        print(f"   Metadata: {result['metadata']}")
    print()

For production applications, embeddings can be stored in any vector database. Amazon OpenSearch Service offers native integration with Nova Multimodal Embeddings at launch, making it straightforward to build scalable search applications. As shown in the examples before, Amazon S3 Vectors provides a simple way to store and query embeddings with your application data.

Things to know
Nova Multimodal Embeddings offers four output dimension options: 3,072, 1,024, 384, and 256. Larger dimensions provide more detailed representations but require more storage and computation. Smaller dimensions offer a practical balance between retrieval performance and resource efficiency. This flexibility helps you optimize for your specific application and cost requirements.

The model handles substantial context lengths. For text inputs, it can process up to 8,192 tokens at once. Video and audio inputs support segments of up to 30 seconds, and the model can segment longer files. This segmentation capability is particularly useful when working with large media files—the model splits them into manageable pieces and creates embeddings for each segment.

The model includes responsible AI features built into Amazon Bedrock. Content submitted for embedding goes through Amazon Bedrock content safety filters, and the model includes fairness measures to reduce bias.

As described in the code examples, the model can be invoked through both synchronous and asynchronous APIs. The synchronous API works well for real-time applications where you need immediate responses, such as processing user queries in a search interface. The asynchronous API handles latency insensitive workloads more efficiently, making it suitable for processing large content such as videos.

Availability and pricing
Amazon Nova Multimodal Embeddings is available today in Amazon Bedrock in the US East (N. Virginia) AWS Region. For detailed pricing information, visit the Amazon Bedrock pricing page.

To learn more, see the Amazon Nova User Guide for comprehensive documentation and the Amazon Nova model cookbook on GitHub for practical code examples.

If you’re using an AI–powered assistant for software development such as Amazon Q Developer or Kiro, you can set up the AWS API MCP Server to help the AI assistants interact with AWS services and resources and the AWS Knowledge MCP Server to provide up-to-date documentation, code samples, knowledge about the regional availability of AWS APIs and CloudFormation resources.

Start building multimodal AI-powered applications with Nova Multimodal Embeddings today, and share your feedback through AWS re:Post for Amazon Bedrock or your usual AWS Support contacts.

— Danilo

This post was originally published on this site

Today, we’re announcing AWS RTB Fabric, a fully managed service purpose built for real-time bidding (RTB) advertising workloads. The service helps advertising technology (AdTech) companies seamlessly connect with their supply and demand partners, such as Amazon Ads, GumGum, Kargo, MobileFuse, Sovrn, TripleLift, Viant, Yieldmo and more, to run high-volume, latency-sensitive RTB workloads on Amazon Web Services (AWS) with consistent single-digit millisecond performance and up to 80% lower networking costs compared to standard networking costs.

AWS RTB Fabric provides a dedicated, high-performance network environment for RTB workloads and partner integrations without requiring colocated, on-premises infrastructure or upfront commitments. The following diagram shows the high-level architecture of RTB Fabric.

AWS RTB Fabric also includes modules, a capability that helps customers bring their own and partner applications securely into the compute environment used for real-time bidding. Modules support containerized applications and foundation models (FMs) that can enhance transaction efficiency and bidding effectiveness. At launch, AWS RTB Fabric includes modules for optimizing traffic management, improving bid efficiency, and increasing bid response rates, all running inline within the service for consistent low-latency execution.

The growth of programmatic advertising has created a need for low-latency, cost-efficient infrastructure to support RTB workloads. AdTech companies process millions of bid requests per second across publishers, supply-side platforms (SSPs), and demand-side platforms (DSPs). These workloads are highly sensitive to latency because most RTB auctions must complete within 200–300 milliseconds and require reliable, high-speed exchange of OpenRTB requests and responses among multiple partners. Many companies have addressed this by deploying infrastructure in colocation data centers near key partners, which reduces latency but adds operational complexity, long provisioning cycles, and high costs. Others have turned to cloud infrastructure to gain elasticity and scale, but they often face complex provisioning, partner-specific connectivity, and long-term commitments to achieve cost efficiency. These gaps add operational overhead and limit agility. AWS RTB Fabric solves these challenges by providing a managed private network built for RTB workloads that delivers consistent performance, simplifies partner onboarding, and achieves predictable cost efficiency without the burden of maintaining colocation or custom networking setups.

Key capabilities
AWS RTB Fabric introduces a managed foundation for running RTB workloads at scale. The service provides the following key capabilities:

• Simplified connectivity to AdTech partners – When you register an RTB Fabric gateway, the service automatically generates secure endpoints that can be shared with selected partners. Using the AWS RTB Fabric API, you can create optimized, private connections to exchange RTB traffic securely across different environments. External Links are also available to connect with partners who aren’t using RTB Fabric, such as those operating on premises or in third-party cloud environments. This approach shortens integration time and simplifies collaboration among AdTech participants.
• Dedicated network for low-latency advertising transactions – AWS RTB Fabric provides a managed, high-performance network layer optimized for OpenRTB communication. It connects AdTech participants such as SSPs, DSPs, and publishers through private, high-speed links that deliver consistent single-digit millisecond latency. The service automatically optimizes routing paths to maintain predictable performance and reduce networking costs, without requiring manual peering or configuration.
• Pricing model aligned with RTB economics – AWS RTB Fabric uses a transaction-based pricing model designed to align with programmatic advertising economics. Customers are billed per billion transactions, providing predictable infrastructure costs that align with how advertising exchanges, SSPs, and DSPs operate.
• Built-in traffic management modules – AWS RTB Fabric includes configurable modules that help AdTech workloads operate efficiently and reliably. Modules such as Rate Limiter, OpenRTB Filter, and Error Masking help you control request volume, validate message formats, and manage response handling directly in the network path. These modules execute inline within the AWS RTB Fabric environment, maintaining network-speed performance without adding application-level latency. All configurations are managed through the AWS RTB Fabric API, so you can define and update rules programmatically as your workloads scale.

Getting started
Today, you can start building with AWS RTB Fabric using the AWS Management Console, AWS Command Line Interface (AWS CLI), or infrastructure-as-code (IaC) tools such as AWS CloudFormation and Terraform.

The console provides a visual entry point to view and manage RTB gateways and links, as shown on the Dashboard of the AWS RTB Fabric console.

You can also use the AWS CLI to configure gateways, create links, and manage traffic programmatically. When I started building with AWS RTB Fabric, I used the AWS CLI to configure everything from gateway creation to link setup and traffic monitoring. The setup ran inside my Amazon Virtual Private Cloud (Amazon VPC) endpoint while AWS managed the low-latency infrastructure that connected workloads.

To begin, I created a requester gateway to send bid requests and a responder gateway to receive and process bid responses. These gateways act as secure communication points within the AWS RTB Fabric.

# Create a requester gateway with required parameters
aws rtbfabric create-requester-gateway 
  --description "My RTB requester gateway" 
  --vpc-id vpc-12345678 
  --subnet-ids subnet-abc12345 subnet-def67890 
  --security-group-ids sg-12345678 
  --client-token "unique-client-token-123"

# Create a responder gateway with required parameters
aws rtbfabric create-responder-gateway 
  --description "My RTB responder gateway" 
  --vpc-id vpc-01f345ad6524a6d7 
  --subnet-ids subnet-abc12345 subnet-def67890 
  --security-group-ids sg-12345678 
  --dns-name responder.example.com 
  --port 443 
  --protocol HTTPS

After both gateways were active, I created a link from the requester to the responder to establish a private, low-latency communication path for OpenRTB traffic. The link handled routing and load balancing automatically.

# Requester account creating a link from requester gateway to a responder gateway
aws rtbfabric create-link 
  --gateway-id rtb-gw-requester123 
  --peer-gateway-id rtb-gw-responder456 
  --log-settings '{"applicationLogs:{"sampling":"errorLog":10.0,"filterLog":10.0}}'

# Responder account accepting a link from requester gateway to responder gateway
aws rtbfabfic accept-link 
  --gateway-id rtb-gw-responder456 
  --link-id link-reqtoresplink789 
  --log-settings '{"applicationLogs:{"sampling":"errorLog":10.0,"filterLog":10.0}}'

I also connected with external partners using External Links, which extended my RTB workloads to on-premises or third-party environments while maintaining the same latency and security characteristics.

# Create an inbound external link endpoint for an external partner to send bid requests to
aws rtbfabric create-inbound-external-link 
  --gateway-id rtb-gw-responder456

# Create an outbound external link for sending bid requests to an external partner
aws rtbfabric create-outbound-external-link 
  --gateway-id rtb-gw-requester123 
  --public-endpoint "https://my-external-partner-responder.com"

To manage traffic efficiently, I added modules directly into the data path. The Rate Limiter module controlled request volume, and the OpenRTB Filter validated message formats inline at network speed.

# Attach a rate limiting module
aws rtbfabric update-link-module-flow 
  --gateway-id rtb-gw-responder456 
  --link-id link-toresponder789 
  --modules '{"name":"RateLimiter":"moduleParameters":{"rateLimiter":{"tps":10000}}}'

Finally, I used Amazon CloudWatch to monitor throughput, latency, and module performance, and I exported logs to Amazon Simple Storage Service (Amazon S3) for auditing and optimization.

All configurations can also be automated with AWS CloudFormation or Terraform, allowing consistent, repeatable deployment across multiple environments. With RTB Fabric, I could focus on optimizing bidding logic while AWS maintained predictable, single-digit millisecond performance across my AdTech partners.

For more details, refer to the AWS RTB Fabric User Guide.

Now available
AWS RTB Fabric is available today in the following AWS Regions: US East (N. Virginia), US West (Oregon), Asia Pacific (Singapore), Asia Pacific (Tokyo), Europe (Frankfurt), and Europe (Ireland).

AWS RTB Fabric is continually evolving to address the changing needs of the AdTech industry. The service expands its capabilities to support secure integration of advanced applications and AI-driven optimizations in real-time bidding workflows that help customers simplify operations and improve performance on AWS. To learn more about AWS RTB Fabric, visit the AWS RTB Fabric page.

– Betty

This post was originally published on this site

Since it launched in 2022, the Customer Carbon Footprint Tool (CCFT) has supported our customers’ sustainability journey to track, measure, and review their carbon emissions by providing the estimated carbon emissions associated with their usage of Amazon Web Services (AWS) services.

In April, we made major updates in the CCFT, including easier access to carbon emissions data, visibility into emissions by AWS Region, inclusion of location-based emissions (LBM), an updated, independently-verified methodology as well as moving to a dedicated page in the AWS Billing console.

The CCFT is informed by the Greenhouse Gas (GHG) Protocol’s classification of emissions, which classifies a company’s emissions. Today, we’re announcing the inclusion of Scope 3 emissions data and an update to Scope 1 emissions in the CCFT. The new emission categories complement the existing Scope 1 and 2 data, and they’ll give our customers a comprehensive look into their carbon emissions data.

In this updated methodology we incorporate new emissions categories. We’ve added Scope 1 refrigerants and natural gas, alongside the existing Scope 1 emissions from fuel combustion in emergency backup generators (diesel). Although Scope 1 emissions represent a small share of overall emissions, we provide our customers with a complete image of their carbon emissions.

To decide which categories of Scope 3 to include in our model we looked at how material each of them were to the overall carbon impact and confirmed the vast majority of emissions were represented. With that in mind, the methodology now includes:

• Fuel- and energy-related activities (“FERA” under the GHG Protocol) – This includes upstream emissions from purchased fuels, upstream emissions of purchased electricity, and transmission and distribution (T&D) losses. AWS calculates these emissions using both LBM and the market-based method (MBM).

• IT hardware – AWS uses a comprehensive cradle-to-gate approach that tracks emissions from raw material extraction through manufacturing and transportation to AWS data centers. We use four calculation pathways: process-based life cycle assessment (LCA) with engineering attributes, extrapolation, representative category average LCA, and economic input-output LCA. AWS prioritizes the most detailed and accurate methods for components that contribute significantly to overall emissions.

• Buildings and equipment – AWS follows established whole building life cycle assessment (wbLCA) standards, considering emissions from construction, use, and end-of-life phases. The analysis covers data center shells, rooms, and long-lead equipment such as air handling units and generators. The methodology uses both process-based life cycle assessment models and economic input-output analysis to provide comprehensive coverage.

The Scope 3 emissions are then amortized over the assets’ service life (6 years for IT hardware, 50 years for buildings) to calculate monthly emissions that can be allocated to customers. This amortization means that we fairly distribute the total embodied carbon of each asset across its operational lifetime, accounting for scenarios such as early retirement or extended use.

All these updates are part of methodology version 3.0.0 and are explained in detail in our methodology document, which has been independently verified by a third party.

How to access the CCFT
To get started, go to the AWS Billing and Cost Management console and choose Customer Carbon Footprint Tool under Cost and Usage Analysis. You can access your carbon emissions data in the dashboard, download a csv file, or export all data using basic SQL and visualize your data by integrating with AWS Data Exports and Amazon Quick Sight.

To ensure you can make meaningful year-over-year comparisons, we’ve recalculated historical data back to January 2022 using version 3 of the methodology. All the data displayed in the CCFT now uses version 3. To see historical data using v3, choose Create custom data export. A new data export now includes new columns breaking down emissions by Scope 1, 2, and 3.

You can see estimated AWS emissions and estimated emissions savings. The tool shows emissions calculated using the MBM for 38 months of data by default. You can find your emissions calculated using the LBM by choosing LBM in the Calculation method filter on the dashboard. The unit of measurement for carbon emissions is metric tons of carbon dioxide equivalent (MTCO2e), an industry-standard measure.

In the Carbon emissions summary, it shows trends of your carbon emissions over time. You can also find emissions resulting from your usage of AWS services and across all AWS Regions. To learn more, visit Viewing your carbon footprint in the AWS documentation.

Voice of the customer
Some of our customers had early access to these updates. This is what they shared with us:

Sunya Norman, senior vice president, Impact at Salesforce shared “Effective decarbonization begins with visibility into our carbon footprint, especially in Scope 3 emissions. Industry averages are only a starting point. The granular carbon data we get from cloud providers like AWS are critical to helping us better understand the actual emissions associated with our cloud infrastructure and focus reductions where they matter most.”

Gerhard Loske, Head of Environmental Management at SAP said “The latest updates to the CCFT are a big step forward in helping us managing SAP’s sustainability goals. With new Region-specific data, we can now see better where emissions are coming from and take targeted action. The upcoming addition of Scope 3 emissions will give us a much fuller picture of our carbon footprint across AWS workloads. These improvements make it easier for us to turn data into meaningful climate action.”

Pinterest’s Global Sustainability Lead, Mia Ketterling highlighted the benefits of the Scope 3 emission data, saying, “By including Scope 3 emissions data in their CCFT, AWS empowers customers like Pinterest to more accurately measure and report the full carbon footprint of our digital operations. Enhanced transparency helps us drive meaningful climate action across our value chain.”

If you’re attending AWS re:Invent in person in December, join technical leaders from AWS, Adobe, and Salesforce as they reveal how the Customer Carbon Footprint Tool supports their environmental initiatives.

Now available
With Scope 1, 2, and 3 coverage in the CCFT, you can track your emissions over time to understand how you’re trending towards your sustainability goals and see the impact of any carbon reduction projects you’ve implemented. To learn more, visit the Customer Carbon Footprint Tool (CCFT) page.

Give these new features a try in the AWS Billing and Cost Management console and send feedback to AWS re:Post for the CCFT or through your usual AWS Support contacts.

— Channy