n8n-nodes-opensearch
v0.2.5
Published
Adds OpenSearch vector store and generic nodes to n8n
Maintainers
Keywords
Readme
n8n-nodes-opensearch
This is an n8n community node. It lets you use OpenSearch in your n8n workflows.
The OpenSearch project, created by Amazon, is a forked search project based on old versions of Elasticsearch and Kibana.
n8n is a fair-code licensed workflow automation platform.
Installation Operations Credentials Compatibility Sample Workflows Version History Resources
Installation
Follow the installation guide in the n8n community nodes documentation.
Operations
Document Actions
- Create a document
- Delete a document
- Get a document
- Get many documents
- Update a document
Index Actions
- Create an index
- Delete an index
- Get an index
- Get many indices
OpenSearch Vector Store
- Retrieve
- Insert
- Get Many
Credentials
You will need a baseURL and a username and password to authenticate to the OpenSearch service.
Compatibility
| Package Version | n8n Version | Notes | |-----------------|-------------|-------| | @0.1.4 | 1.48.0+ | Tested with 1.62.1 | | @0.2.x | 2.0.0+ | Full support including AI tools | | @0.2.x | 1.100.0+ | AI tools supported |
Using as AI Tool
The OpenSearch node supports usableAsTool: true, which allows it to be used as an AI Agent tool. This creates an opensearchTool node type automatically.
Requirements:
- n8n 1.100.0 or later is required for AI tool support with community packages. Earlier versions of n8n 1.x (< 1.100.0) only support tool nodes from built-in
n8n-nodes-basepackages.
Query Parameter: The Search Index operation's Query parameter accepts:
- Plain text: e.g.,
"my search term"- automatically converted to aquery_stringsearch across all fields - JSON: e.g.,
{"query": {"match": {"title": "hello"}}}- passed directly to OpenSearch
When using as an AI tool, set up the Query field with $fromAI():
{{ $fromAI('Query', 'Search term to find documents', 'string') }}Important: The $fromAI() must use 'string' type (not 'json'). If upgrading from an older version where you had 'json' type, you'll need to either:
- Manually change
'json'to'string'in the expression - Delete and re-add the tool node, then click "let the model define the parameters"
n8n 1.x vs 2.x: Both versions work with AI tools. n8n 1.x uses AgentV2 which validates tool inputs with Zod schemas directly. n8n 2.x uses AgentV3 with a different execution model. Using 'string' type ensures compatibility with both.
Sample Workflows
Sample n8n workflows demonstrating OpenSearch node usage are available in the tests/workflows/ folder:
opensearch-all-modes-test-1.123.7.json- For n8n 1.xopensearch-all-modes-test-2.1.0.json- For n8n 2.x
Import these into your n8n instance to see examples of all operations including document CRUD, index management, and vector store usage.
Version History
0.2.5 (2025-12-19)
- Fixed PostgreSQL/MySQL compatibility for VectorStoreOpenSearch node versioning
- Migrated to
@n8n/node-clifor build tooling - Added dark icon variants for theme support
- Improved UX: cleaner operation names, better descriptions, proper delete output format
- Moved codex metadata to separate
.node.jsonfiles (scaffold pattern)
0.2.4 (2025-12-18)
- Fixed module resolution issue preventing installation via n8n community package installer
- Replaced
@langchain/communityOpenSearch vectorstore with custom implementation to avoid dependency conflicts
0.2.3 (2025-12-18)
- Moved
@langchain/communityand@langchain/coreto peerDependencies to avoid version conflicts
0.2.2 (2025-12-18)
- Fixed OpenSearch node not appearing in n8n 2.x search (removed incorrect 'AI' category)
0.2.1 (2025-12-18)
- Extensive refactor with full test coverage
- Added support for OpenSearch 2.x and 3.x
- Added nmslib engine deprecation handling for OpenSearch 3.x
- Fixed build and packaging issues
0.2.0 (2025-12-17)
- Upgraded to
@langchain/classic1.0.5 for n8n 2.x compatibility - Added n8n 2.x support while maintaining n8n 1.x compatibility
0.1.3 (2025-10-02)
- Added Search Index operation with Scroll API support
- Fixed HTTP methods for complex queries
0.1.2 (2025-07-10)
- Fixed missing dependency issue
0.1.1 (2024-10-04)
- Documentation updates
0.1.0 (2024-10-04)
- Initial release with OpenSearch document and index operations
- OpenSearch Vector Store for AI/LangChain workflows
