- Repository introduced with following examples:
- Hello World Agent
- DevOps Agent
- Energy Efficiency Management Agent
- Portfolio Assistant Agent
- Startup Advisor Agent
- Sports Team Poet Agent
- Trip Planner Agent
- Voyage Vituoso Agent
- Repository introduced with following shared tool. Deployment is only supported with CloudFormation Stack:
- Repository introduced with bedrock_agent_helper, bedrock_agent, and knowledge_base_helper.
Renaming examples folder for simplicity and adding documentation for single agents examples
- Following examples added for Amazon Bedrock Agents:
- Analyst assistant using Code Interpretation
- Agent using Amazon Bedrock Guardrails
- Agent using Amazon Bedrock Knowledge Bases
- Agent with long term memory
- Agent using models not yet optimized for Bedrock Agents
- AWS CDK Agent
- Custom orchestration Agent
- Configure an inline agent at runtime
- Utilize LangChain Tools with Amazon Bedrock Inline Agents
- Provide conversation history to Amazon Bedrock Agents
- Agent using OpenAPI schema
- Agents with user confirmation before action execution
- Streamlit UI demo added.
- Following extensions added to bedrock_agent library
- Added helper methods
delete_by_name()
andexists()
toAgent
to allow usingbedrock_agent
for common operations without needing to usebedrock_agent_helper
directly - Reworked
Tool
class - Added
attach_tool()
toAgent
to allow attaching tools after creation - Added
attach_tool_from_function()
to Agent to create a tool from Python code with type hints (note it currently materializes a lambda from the function) - Added explicit
update()
andprepare()
methods toAgent
- Added
create_from_yaml()
toAgent
, allows creating fom a definition in a YAML file
- Following documentation added for Amazon Bedrock Agents:
- Added Getting Started tutorial for using the bedrock_agent library with basic usage examples and coverage of the new functions added below
- All examples in multi_agent_collaboration that used both
bedrock_agent
andbedrock_agent_helper
have been updated to just usebedrock_agent
withAgent
helper functions
- Following usability improvements made to bedrock_agent library:
- Renamed
direct_create()
inAgent
andSupervisorAgent
tocreate()
for simplicity - When creating from YAML,
Agent
will automatically de-indent (dedent) the description - Agent will now automatically
prepare()
internally if needed wheninvoke()
, removing a step from user code
- Following extensions added to bedrock_agent_helper library:
- Added
create_lambda_file()
, which will create aa Lambda function from supplied Python code with type hints
Added to Agent using models not yet optimized for Bedrock Agents By creating a code sample for llama as an unoptimized model
Adding Agent with access to house security camera in cloudformation