Strands Agents

​

What is Strands Agents?

​

Key Features of Strands Agents

• Model-Driven Orchestration: Leverages model reasoning to plan, orchestrate tasks, and reflect on goals autonomously
• Model & Provider Agnostic: Work with any LLM provider - Amazon Bedrock, OpenAI, Anthropic, local models - without changing your code
• Multi-Agent Primitives: Simple primitives for handoffs, swarms, and graph workflows with built-in support for Agent-to-Agent (A2A) communication
• Native AWS Integration: Best-in-class AWS integrations with easy deployment to EKS, Lambda, EC2, and native MCP tool integration

​

How TrueFoundry Integrates with Strands Agents

​

Installation & Setup

1

Install Strands Agents

pip install -U strands-agents strands-agents-tools

2

Get TrueFoundry Access Token

1. Sign up for a TrueFoundry account
2. Follow the steps here in Quick start

3

Configure Strands with TrueFoundry

from strands.models.openai import OpenAIModel

# Create a model instance with TrueFoundry AI Gateway
truefoundry_model = OpenAIModel(
    client_args={
        "api_key": "your_truefoundry_api_key",
        "base_url": "{GATEWAY_BASE_URL}"
    },
    model_id="openai-main/gpt-4o",  
    params={"temperature": 0.7}
)

# Use in your Strands agent
from strands import Agent

agent = Agent(model=truefoundry_model)
response = agent("What is 2+2?")

​

Multi-Provider Support

# OpenAI
openai_model = OpenAIModel(
    client_args={
        "api_key": "your_truefoundry_api_key",
        "base_url": "{GATEWAY_BASE_URL}"
    },
    model_id="llm-gateway-prod/gpt-4o",
    params={"temperature": 0.7}
)

# Anthropic Claude
claude_model = OpenAIModel(
    client_args={
        "api_key": "your_truefoundry_api_key",
        "base_url": "{GATEWAY_BASE_URL}"
    },
    model_id="llm-gateway-prod/claude-sonnet-4-5",
    params={"temperature": 0.7}
)

# Google Gemini
gemini_model = OpenAIModel(
    client_args={
        "api_key": "your_truefoundry_api_key",
        "base_url": "{GATEWAY_BASE_URL}"
    },
    model_id="llm-gateway-prod/gemini-2-0-flash-lite-001",
    params={"temperature": 0.7}
)

# Use different models for different agents
researcher = Agent(model=claude_model, tools=[web_search])
calculator_agent = Agent(model=openai_model, tools=[calculator])

​

Observability and Governance

• Performance Metrics: Track key latency metrics like Request Latency, Time to First Token (TTFS), and Inter-Token Latency (ITL) with P99, P90, and P50 percentiles
• Cost and Token Usage: Gain visibility into your application’s costs with detailed breakdowns of input/output tokens and the associated expenses for each model
• Usage Patterns: Understand how your application is being used with detailed analytics on user activity, model distribution, and team-based usage
• Rate Limiting and Virtual Models: Set up rate limiting and configure Virtual Models for intelligent routing and fallback across your models