What is Agent Builder

Agent-builder is a command-line interface (CLI) assistant powered by Strands Agents, designed to facilitate the creation of interactive agents and tools that can perform various tasks, including streaming and tool usage.

Use cases

Use cases for agent-builder include summarizing documents, executing Python code, making API calls, managing environment variables, and creating interactive dialog interfaces for user engagement.

How to use

To use agent-builder, install it via pipx with the command ‘pipx install strands-agents-builder’. You can run it in interactive mode by typing ‘strands’, or make one-off queries by passing arguments or piping content into it.

Key features

Key features include an interactive command-line interface, one-off query capabilities, integration of over 12 powerful tools (such as shell, editor, and HTTP), knowledge base integration for context and memory, customizable prompts, nested agent capabilities, dynamic tool loading, and environment variable management.

Where to use

Agent-builder can be used in various fields such as software development, data analysis, project management, and any domain where interactive command-line tools can enhance productivity and efficiency.

Strands Agent Builder

A model-driven approach to building AI agents in just a few lines of code.

Documentation ◆ Samples ◆ Python SDK ◆ Tools ◆ Agent Builder ◆ MCP Server

An interactive Strands agent toolkit designed to help you build, test, and extend your own custom AI agents and tools. With the Strands Agent Builder, you can create specialized agents, develop custom tools, and compose complex AI workflows—all from your terminal.

Quick Start

# Install
pipx install strands-agents-builder

# Run interactive mode for agent development
strands

# Build a custom tool and use it immediately
strands "Create a tool named sentiment_analyzer that analyzes text sentiment and test it with some examples"

# Pipe content to build an agent based on specifications
cat agent-spec.txt | strands "Build a specialized agent based on these specifications"

# Use with knowledge base to extend existing tools
strands --kb YOUR_KB_ID "Load my previous calculator tool and enhance it with scientific functions"

Features

• 🏗️ Create and test custom tools with instant hot-reloading
• 🤖 Build specialized agents with focused capabilities
• 🔄 Extend existing tools and enhance their functionality
• 💬 Interactive command-line interface with rich output
• 🛠️ Powerful integrated tools (12+ tools including shell, editor, HTTP, Python)
• 🧠 Knowledge base integration for persisting and loading tools
• 🎮 Customizable system prompt for specialized agents
• 🪄 Nested agent capabilities with tool delegation
• 🔧 Dynamic tool loading for extending functionality
• 🖥️ Environment variable management and customization

Integrated Tools

Strands comes with a comprehensive set of built-in tools:

• agent_graph: Create and manage graphs of agents
• calculator: Perform mathematical operations
• cron: Task scheduling with cron jobs (not available on Windows)
• current_time: Get the current date and time
• editor: File editing operations like line edits, search, and undo
• environment: Manage environment variables
• generate_image: Create AI generated images with Amazon Bedrock
• http_request: Make API calls, fetch web data, and call local HTTP servers
• image_reader: Process and analyze images
• journal: Create structured tasks and logs for agents to manage and work from
• load_tool: Dynamically load more tools at runtime
• memory: Agent memory persistence in Amazon Bedrock Knowledge Bases
• nova_reels: Create AI generated videos with Nova Reels on Amazon Bedrock
• python_repl: Run Python code (not available on Windows)
• retrieve: Semantically retrieve data from Amazon Bedrock Knowledge Bases for RAG, memory, and other purposes
• shell: Execute shell commands (not available on Windows)
• slack: Slack integration with real-time events, API access, and message sending
• speak: Generate speech from text using macOS say command or Amazon Polly
• stop: Force stop the agent event loop
• store_in_kb: Save content to knowledge bases for future reference
• strand: Create nested agent instances with specialized capabilities
• swarm: Coordinate multiple AI agents in a swarm / network of agents
• think: Perform deep thinking by creating parallel branches of agentic reasoning
• use_aws: Interact with AWS services
• use_llm: Run a new AI event loop with custom prompts
• welcome: Manage the Strands Agent Builder welcome text
• workflow: Orchestrate sequenced workflows

Knowledge Base Integration

Strands Agent Builder leverages Amazon Bedrock Knowledge Bases to store and retrieve custom tools, agent configurations, and development history.

# Load and extend tools from your knowledge base
strands --kb YOUR_KB_ID "Load my data_visualizer tool and add 3D plotting capabilities"

# Or set a default knowledge base via environment variable
export STRANDS_KNOWLEDGE_BASE_ID="YOUR_KB_ID"
strands "Find my most recent agent configuration and make it more efficient"

Features:

• 🔄 Retrieve previously created tools and agent configurations
• 💾 Persistent storage for your custom tools and agents
• 🛠️ Ability to iteratively improve tools across sessions
• 🔍 Find and extend tools built in previous sessions

Model Configuration

Optimized Defaults

Strands comes with optimized, maxed-out configuration settings for the Bedrock model provider:

{
  "model_id": "us.anthropic.claude-sonnet-4-20250514-v1:0",
  "max_tokens": 32767,
  "boto_client_config": {
    "read_timeout": 900,
    "connect_timeout": 900,
    "retries": {
      "max_attempts": 3,
      "mode": "adaptive"
    }
  },
  "additional_request_fields": {
    "anthropic_beta": [
      "interleaved-thinking-2025-05-14"
    ],
    "thinking": {
      "type": "enabled",
      "budget_tokens": 2048
    }
  }
}

These settings provide:

• Claude Sonnet 4 (latest high-performance model)
• Maximum token output (32,768 tokens)
• Extended timeouts (15 minutes) for complex operations
• Automatic retries with adaptive backoff
• Interleaved thinking capability for real-time reasoning during responses
• Enabled thinking capability with 2,048 token budget for recursive reasoning

You can customize these values using environment variables:

# Maximum tokens for responses
export STRANDS_MAX_TOKENS=32000

# Budget for agent thinking/reasoning
export STRANDS_BUDGET_TOKENS=1024

Custom Model Provider

You can configure strands to use a different model provider with specific settings by passing in the following arguments:

strands --model-provider <NAME> --model-config <JSON|FILE>

As an example, if you wanted to use the packaged Ollama provider with a specific model id, you would run:

strands --model-provider ollama --model-config '{"model_id": <ID>}'

Strands Agent Builder is packaged with bedrock and ollama.

If you have implemented a custom model provider (instructions) and would like to use it with strands, create a python module under the directory “$CWD/.models” and expose an instance function that returns an instance of your provider. As an example, assume you have:

$ cat ./.models/custom_model.py
from mymodels import CustomModel

def instance(**config):
    return CustomModel(**config)

You can then use it with strands by running:

$ strands --model-provider custom_model --model-config <JSON|FILE>

Custom System Prompts

# Via environment variable
export STRANDS_SYSTEM_PROMPT="You are a Python expert."

# Or local file
echo "You are a security expert." > .prompt

🌍 Environment Variables Configuration

Strands Agent Builder also provides customization through environment variables:

Exit

Type exit, quit, or press Ctrl+C/Ctrl+D

Contributing ❤️

We welcome contributions! See our Contributing Guide for details on:

• Reporting bugs & features
• Development setup
• Contributing via Pull Requests
• Code of Conduct
• Reporting of security issues

License

This project is licensed under the Apache License 2.0 - see the LICENSE file for details.

Security

See CONTRIBUTING for more information.

⚠️ Preview Status

Strands Agents is currently in public preview. During this period:

• APIs may change as we refine the SDK
• We welcome feedback and contributions