Semantic Search
Natural language search over any content
About
Replace keyword search with meaning-based search powered by vector embeddings. Ingest any text content — articles, docs, product descriptions — and query with natural language. The system understands intent, not just keywords.
Supports multi-query search for complex information needs and returns relevance-scored results with metadata.
API Endpoints
/ingest/search/search/multi/docs/healthHow It Works
Content Ingestion
POST /ingest — text content chunked, embedded, and indexed in the vector database.
Query Embedding
Search query converted to a vector in the same embedding space as your content.
Vector Search
Top-k nearest neighbors retrieved with relevance scores (configurable threshold).
Result Ranking
Results ranked by cosine similarity and returned with metadata and scores.
Use Cases
Documentation Search
Let users search your docs with natural language instead of exact keyword matches.
Product Catalog Search
Find products by describing what you need rather than knowing exact names or SKUs.
Research Assistant
Search across a corpus of papers, articles, or reports by meaning.
Content Discovery
Surface related content across your site or app based on semantic similarity.
Opens Aerostack dashboard to deploy this template
What's Included
Pipeline
Billing Model
metered
Pay per token used. Free tier included.