Skip to content

Adding Checkpointer and Store Implementation of Amazon Bedrock AgentCore Memory #1935

Open
hasnainvirk wants to merge 4 commits intolangchain-ai:mainfrom
hasnainvirk:dev
Open

Adding Checkpointer and Store Implementation of Amazon Bedrock AgentCore Memory #1935
hasnainvirk wants to merge 4 commits intolangchain-ai:mainfrom
hasnainvirk:dev

Conversation

@hasnainvirk
Copy link

Implements both BaseCheckpointSaver and BaseStore for AWS Bedrock AgentCore Memory:

- Full checkpoint persistence with thread isolation via sessionId
- Supports all standard operations: getTuple, list, put, putWrites, deleteThread
- Handles checkpoint_ns filtering and metadata search
- Automatic retry logic with exponential backoff for AWS API throttling
- Rate limiting (16.7 req/sec) to stay within AgentCore Memory limits
- Unique actor ID generation for test isolation
- Proper Unicode handling with TextEncoder/TextDecoder

- Hierarchical namespace organization for data isolation
- Metadata filtering and pagination for search operations
- Complex JSON value support with proper serialization
- Vector similarity search capabilities via RetrieveMemoryRecordsCommand
- Consistent rate limiting and error handling

- thread_id maps directly to sessionId in AgentCore Memory
- actor_id provides user-level isolation across sessions
- checkpoint_ns stored in event payload for namespace filtering
- Store namespaces map to sessionId/actorId for hierarchical organization

- Proper TypeScript interfaces for all AWS SDK interactions
- No explicit 'any' types - uses proper type definitions
- Comprehensive error handling with typed AWS error interfaces

@changeset-bot
Copy link

changeset-bot bot commented Feb 2, 2026

🦋 Changeset detected

Latest commit: 963bb02

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 2 packages
Name Type
@langchain/langgraph-checkpoint-aws-agentcore-memory Minor
@langchain/langgraph-checkpoint-validation Minor

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

Hasnain Virk added 3 commits February 3, 2026 10:48
- Add AgentCore Memory validation test initializer
- Export deleteThreadTests for comprehensive testing
- Fix type export for CheckpointSaverTestInitializer
- Increase test timeouts for AWS API throttling (2 minutes)
- Add CLI build support for validation runner

These changes enable validation testing of AWS-based checkpoint implementations
that may experience API throttling and require longer test execution times.
Implements both BaseCheckpointSaver and BaseStore for AWS Bedrock AgentCore Memory:

- Full checkpoint persistence with thread isolation via sessionId
- Supports all standard operations: getTuple, list, put, putWrites, deleteThread
- Handles checkpoint_ns filtering and metadata search
- Automatic retry logic with exponential backoff for AWS API throttling
- Rate limiting (16.7 req/sec) to stay within AgentCore Memory limits
- Unique actor ID generation for test isolation
- Proper Unicode handling with TextEncoder/TextDecoder

- Hierarchical namespace organization for data isolation
- Metadata filtering and pagination for search operations
- Complex JSON value support with proper serialization
- Vector similarity search capabilities via RetrieveMemoryRecordsCommand
- Consistent rate limiting and error handling

- thread_id maps directly to sessionId in AgentCore Memory
- actor_id provides user-level isolation across sessions
- checkpoint_ns stored in event payload for namespace filtering
- Store namespaces map to sessionId/actorId for hierarchical organization

- Proper TypeScript interfaces for all AWS SDK interactions
- No explicit 'any' types - uses proper type definitions
- Comprehensive error handling with typed AWS error interfaces

Includes comprehensive integration tests and validation test support.
Add AWS SDK dependencies and development tools required for the new
checkpoint-aws-agentcore-memory package.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant