Cloud Platform Integrations

TraceCraft integrates with major cloud AI platforms.

Supported Platforms

Platform	Status	Installation
AWS AgentCore	Stable	`tracecraft[aws-agentcore]`
Azure AI Foundry	Stable	`tracecraft[azure-foundry]`
GCP Vertex Agent Builder	Stable	`tracecraft[gcp-vertex-agent]`
All Platforms	-	`tracecraft[cloud]`

AWS AgentCore

Integration with AWS Bedrock AgentCore.

Installation

pip install "tracecraft[aws-agentcore]"

Configuration

from tracecraft.core.config import TraceCraftConfig, AWSAgentCoreConfig
 
config = TraceCraftConfig(
    aws_agentcore=AWSAgentCoreConfig(
        enabled=True,
        use_xray_propagation=True,
        session_id="conversation-123",
    )
)
 
tracecraft.init(config=config)

Environment Variables

export TRACECRAFT_AWS_AGENTCORE_ENABLED=true
export TRACECRAFT_AWS_XRAY_PROPAGATION=true
export TRACECRAFT_AWS_SESSION_ID=conversation-123
export OTEL_EXPORTER_OTLP_ENDPOINT=http://localhost:4317  # ADOT collector

Features

X-Ray trace propagation
Session tracking
CloudWatch Logs integration
ADOT collector support

See AWS AgentCore Deployment for details.

Azure AI Foundry

Integration with Azure AI Foundry (Project Geneva).

Installation

pip install "tracecraft[azure-foundry]"

Configuration

from tracecraft.core.config import TraceCraftConfig, AzureFoundryConfig
 
config = TraceCraftConfig(
    azure_foundry=AzureFoundryConfig(
        enabled=True,
        connection_string=os.environ["APPLICATIONINSIGHTS_CONNECTION_STRING"],
        enable_content_recording=True,
        agent_name="customer-support",
        agent_id="agent-v1",
    )
)
 
tracecraft.init(config=config)

Environment Variables

export TRACECRAFT_AZURE_FOUNDRY_ENABLED=true
export APPLICATIONINSIGHTS_CONNECTION_STRING="InstrumentationKey=..."
export TRACECRAFT_AZURE_CONTENT_RECORDING=true
export TRACECRAFT_AZURE_AGENT_NAME=customer-support

Features

Application Insights integration
Agent metadata tracking
Content recording (prompts/responses)
Azure Monitor compatibility

See Azure AI Foundry Deployment for details.

GCP Vertex Agent Builder

Integration with Google Cloud Vertex AI Agent Builder.

Installation

pip install "tracecraft[gcp-vertex-agent]"

Configuration

from tracecraft.core.config import TraceCraftConfig, GCPVertexAgentConfig
 
config = TraceCraftConfig(
    gcp_vertex_agent=GCPVertexAgentConfig(
        enabled=True,
        project_id="my-project",
        session_id="session-123",
        agent_name="support-agent",
        enable_content_recording=True,
    )
)
 
tracecraft.init(config=config)

Environment Variables

export TRACECRAFT_GCP_VERTEX_ENABLED=true
export GOOGLE_CLOUD_PROJECT=my-project
export TRACECRAFT_GCP_SESSION_ID=session-123
export TRACECRAFT_GCP_AGENT_NAME=support-agent

Features

Cloud Trace integration
Agent session tracking
Content recording
Vertex AI compatibility

See GCP Vertex Agent Deployment for details.

Multi-Cloud

Use TraceCraft with multiple cloud platforms:

config = TraceCraftConfig(
    # AWS
    aws_agentcore=AWSAgentCoreConfig(enabled=True),
 
    # Azure
    azure_foundry=AzureFoundryConfig(
        enabled=True,
        connection_string=os.environ["APPLICATIONINSIGHTS_CONNECTION_STRING"],
    ),
 
    # GCP
    gcp_vertex_agent=GCPVertexAgentConfig(
        enabled=True,
        project_id=os.environ["GOOGLE_CLOUD_PROJECT"],
    ),
)
 
tracecraft.init(config=config)

Next Steps

OpenTelemetry Receiver Overview