search
This skill guides you through building complete search applications on OpenSearch, covering index setup, document processing, and multiple search strategies including BM25, semantic, hybrid, and agentic approaches. Deploy ML models, ingest PDFs and other documents, and evaluate search quality end-to-end.
Search helps you build and deploy a complete search application with OpenSearch from sample data to a running UI.
AI-generated summary based on this skill's SKILL.md
Decision gist · record as of 2026-07-22
Search helps you build and deploy a complete search application with OpenSearch from sample data to a running UI. This skill guides you through building complete search applications on OpenSearch, covering index setup, document processing, and multiple search strategies including BM25, semantic, hybrid, and agentic approaches. Deploy ML models, ingest PDFs and other documents, and evaluate search quality end-to-end.
Use it when
- OpenSearch skill covers multiple search approaches: BM25 for keyword-based full-text search, semantic search using embeddings and ML models.
- OpenSearch skill walks you through semantic search setup by deploying ML models to generate embeddings.
Verify before relying
Read SKILL.md below before installing (9 files). Open directory: indexed for reading, not audited.
Similar skills
Install
opensearch-project/opensearch-agent-skills/search · repository language: Python
Open directory. Skills are indexed for reading, not audited. Review a skill's body before installing it.
Frequently asked questions
AI-generated answers based on this skill's SKILL.md and metadata
How do I build a search app with OpenSearch?
OpenSearch skill guides you through building complete search applications from scratch. Start by setting up indexes, ingesting your documents (PDFs, text, etc.), and choosing your search strategy—BM25 for traditional full-text, semantic for meaning-based retrieval, or hybrid combining both. Deploy ML models for embeddings, configure your search UI, and iteratively tune relevance. The skill covers the entire end-to-end workflow.
What search strategies does OpenSearch skill support?
OpenSearch skill covers multiple search approaches: BM25 for keyword-based full-text search, semantic search using embeddings and ML models, hybrid search combining sparse and dense vectors, and agentic search strategies. Each approach suits different use cases—choose based on your content type, query patterns, and relevance requirements. The skill helps you implement and evaluate each.
How do I set up semantic or vector search with embeddings?
OpenSearch skill walks you through semantic search setup by deploying ML models to generate embeddings, configuring vector indexes (including KNN search), and implementing dense vector retrieval. You'll learn to ingest documents, transform them into embeddings, and query using semantic similarity. The skill also covers hybrid approaches that combine vector search with traditional BM25.
How can I ingest and index documents like PDFs for search?
OpenSearch skill teaches document processing and ingestion for search applications. You'll learn to extract content from PDFs and other formats, structure documents for indexing, and set up BM25 or semantic indexes. The skill covers the full pipeline from raw documents to searchable indexes, including preprocessing and optimization for retrieval quality.
How do I evaluate and improve search quality in OpenSearch?
OpenSearch skill includes guidance on evaluating and tuning search relevance. You'll learn to measure search quality, adjust ranking parameters, optimize your index configuration, and compare different search strategies. The skill helps you iterate on relevance metrics and deploy improvements to your search application.
What is hybrid search and agentic search in OpenSearch?
OpenSearch skill covers hybrid search, which combines sparse vector (BM25) and dense vector (semantic) approaches for better relevance. Agentic search strategies use AI agents to orchestrate complex retrieval workflows. The skill teaches implementation patterns for both, helping you build sophisticated search systems that adapt to diverse query types and content.
SKILL.md
Rendered from the published skill. Quoted content, verbatim.
Search
Category skill for building search applications with OpenSearch.
Skills
| Skill | Description |
|---|---|
| opensearch-launchpad | End-to-end search application builder — from sample data to a running search UI with BM25, semantic, hybrid, or agentic search |
When to Use
Read opensearch-launchpad/SKILL.md when the user wants to: - Build a new search application - Set up an OpenSearch index with a specific search strategy - Deploy ML models for semantic or hybrid search - Evaluate and tune search quality - Process documents (PDF, DOCX) for search ingestion
File tree — 9 files
skills/opensearch-skills/search/SKILL.md
skills/opensearch-skills/search/opensearch-launchpad/SKILL.md
skills/opensearch-skills/search/opensearch-launchpad/agentic_search_guide.md
skills/opensearch-skills/search/opensearch-launchpad/dense_vector_models.md
skills/opensearch-skills/search/opensearch-launchpad/evaluation_guide.md
skills/opensearch-skills/search/opensearch-launchpad/local_ase.md
skills/opensearch-skills/search/opensearch-launchpad/opensearch_semantic_search_guide.md
skills/opensearch-skills/search/opensearch-launchpad/sparse_vector_models.md
skills/opensearch-skills/search/opensearch-launchpad/unstructured_preset.md
Let your AI agent find skills like this
Example. Real query, live index.
You found this page by searching. An agent finds it by wishing: SkillFed indexes 56,283 agent skills by what they can do, searchable in plain language.
wish › “Build and deploy a complete search application from scratch”
Give your agent the search over MCP, or paste the wish link into any chat. No install? Search from any chat →
Related skills
OpenSearch Skills provides a modular toolkit for constructing search applications with advanced retrieval strategies including vector and semantic search, plus observability features for log analysis and trace investigation. Deploy locally or to AWS, process documents at scale, and route queries intelligently across search, observability, and cloud infrastructure tasks.
OpenSearch Launchpad is your guided path from initial requirements to a fully operational search system. It handles data ingestion, strategy selection (keyword, semantic, hybrid, or agentic), and deployment to either local Docker or Amazon OpenSearch Serverless. The skill walks you through architecture planning and launches a search UI ready for queries.
Provision and configure Amazon OpenSearch Service domains or Serverless collections, then deploy your search configurations to AWS. Handles infrastructure setup, IAM role configuration, Bedrock connector integration, and supports both managed domains and serverless deployments with optional agentic search capabilities.
ingest is a category skill for transforming unstructured files into JSONL chunks on your machine. It handles PDF, DOCX, and PPTX formats through the document-processing skill, which uses Docling to produce search-ready output without requiring cloud infrastructure or AWS.
Scale document ingestion into Amazon OpenSearch using OSIS pipelines. Upload pre-chunked JSONL to S3 for indexing, or send raw PDFs to the cloud for automatic parsing and embedding. Supports both OpenSearch Service domains and Serverless collections.
Document Processing transforms unstructured files into indexed-ready JSONL chunks using the open-source Docling library, running entirely on your machine. Output includes text, headings, source file references, and page numbers for direct ingestion into OpenSearch.
More skills cloud (Apache-2.0) · parallel-web (MIT) · observability (Apache-2.0) · trace-analytics (Apache-2.0)