Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.asteragents.com/llms.txt

Use this file to discover all available pages before exploring further.

​
What it does

The Query Pinecone tool searches vector embeddings in your Pinecone database to find semantically similar documents. Perfect for RAG (Retrieval-Augmented Generation), semantic search, recommendation systems, and finding relevant content based on meaning rather than keywords.
Requires Pinecone Integration: You need to set up a Pinecone integration before agents can use this tool.

​
Key features

  • Semantic search using vector embeddings
  • Search within specific namespaces for organized data
  • Configurable result limits for performance optimization
  • Access to similarity scores and metadata
  • High-performance vector similarity matching

​
Parameters

ParameterTypeRequiredDescription
querystringYesThe search query to find similar documents
namespacestringYesThe namespace within Pinecone to search
top_knumberNoNumber of results to return (default: 5)

​
Common use cases

​
Document retrieval for RAG

query: "How to implement authentication in web applications"
namespace: "documentation"
top_k: 10
Find relevant documentation for answering user questions.

​
Content recommendation

query: "machine learning tutorials for beginners"
namespace: "educational_content"
top_k: 5
Recommend similar content based on user interests. 
query: "customer complaints about product quality"
namespace: "support_tickets"
top_k: 15
Find semantically similar support tickets or feedback.

​
Research assistance

query: "climate change impact on agriculture"
namespace: "research_papers"
top_k: 20
Discover relevant research papers and studies.

​
Product matching

query: "wireless bluetooth headphones with noise cancellation"
namespace: "product_catalog"
top_k: 8
Find similar products in your catalog. 
query: "troubleshooting network connectivity issues"
namespace: "internal_kb"
top_k: 12
Search internal knowledge bases for solutions.

​
Understanding namespaces

Namespaces in Pinecone help organize your vector data:
  • Separate datasets: Keep different types of content isolated
  • Access control: Control which data agents can search
  • Performance: Search within smaller, focused datasets
  • Organization: Logical grouping of related documents
Common namespace patterns:
  • documents - General document storage
  • products - Product catalogs and descriptions
  • support - Support tickets and solutions
  • research - Research papers and studies
  • user_content - User-generated content

​
What you get back

  • Similarity Score: How closely each result matches your query (0-1 scale)
  • Document Metadata: Associated information about each document
  • Match Ranking: Results ordered by relevance/similarity
  • Vector IDs: Unique identifiers for each matched document

​
Best practices

  • Use descriptive, natural language queries for better semantic matching
  • Choose appropriate namespaces to focus your search
  • Adjust top_k based on your use case (more results = broader coverage)
  • Include relevant context in your queries for better matches
  • Monitor similarity scores to understand result quality
  • Organize your Pinecone data with meaningful namespaces

​
Troubleshooting

“Namespace not found”
  • Verify the namespace exists in your Pinecone index
  • Check the namespace spelling (case-sensitive)
  • Ensure the namespace has been populated with data
“No results found”
  • Try broader or different query terms
  • Check if the namespace contains relevant data
  • Increase the top_k value to get more results
“Integration not configured”
  • Set up the Pinecone integration first
  • Verify your API key and host configuration
  • Check that the integration is active and connected
“Low similarity scores”
  • Your query might not match the indexed content well
  • Try rephrasing your query with different terms
  • Check if the right namespace is being searched