MCP Tools Reference

The vai MCP server exposes 11 tools across 5 domains that AI agents can call via the Model Context Protocol.

Retrieval Tools

vai_query

Full RAG query: embed query text, vector search MongoDB, rerank results.

Parameter	Type	Required	Description
query	string	✅	Search query text
db	string	—	Database name (falls back to project config)
collection	string	—	Collection name
model	string	—	Embedding model
limit	number	—	Number of results

vai_search

Vector search without reranking — faster, simpler retrieval.

Parameter	Type	Required	Description
query	string	✅	Search query text
db	string	—	Database name
collection	string	—	Collection name
limit	number	—	Number of results

Embedding Tools

vai_embed

Generate a vector embedding for text.

Parameter	Type	Required	Description
text	string	✅	Text to embed
model	string	—	Embedding model
inputType	string	—	query or document
dimensions	number	—	Output dimensions

vai_similarity

Compare two texts semantically using cosine similarity.

Parameter	Type	Required	Description
text1	string	✅	First text
text2	string	✅	Second text
model	string	—	Embedding model

Management Tools

vai_collections

List collections in a database with vector index information.

Parameter	Type	Required	Description
db	string	—	Database name

vai_models

List available Voyage AI models.

Parameter	Type	Required	Description
category	string	—	Filter: embedding, reranking, or all

Utility Tools

vai_topics

List available educational topics from vai explain.

Parameter	Type	Required	Description
query	string	—	Filter/search topics by keyword

vai_explain

Get a detailed explanation of an embedding/RAG concept.

Parameter	Type	Required	Description
topic	string	✅	Topic key (e.g., embeddings, reranking)

Ingest Tool

vai_ingest

Chunk, embed, and store a document in MongoDB.

Parameter	Type	Required	Description
text	string	✅	Document text to ingest
source	string	—	Source identifier for metadata
db	string	—	Database name
collection	string	—	Collection name
model	string	—	Embedding model
chunkStrategy	string	—	Chunking strategy
chunkSize	number	—	Chunk size in characters

Using MCP Tools

# Install into your AI tool
vai mcp install all

# Start the server manually (for testing)
vai mcp --verbose

Once installed, your AI agent can call these tools directly. For example, in Claude Desktop, you can ask "Search my knowledge base for authentication docs" and Claude will call vai_query automatically.

Further Reading

• vai mcp — MCP server command reference
• MCP Tool Parameters — Detailed parameter schemas
• MCP Server Guide — Setup guide