Embedding Search FAQs

Question 1

How does Embedding Search work?

Accepted Answer

It uses a Turso database containing pre-calculated vector embeddings of transcript segments. When you provide a query, the server calculates the cosine similarity between the query and the stored embeddings, returning the most relevant segments.

Question 2

What are the key features of Embedding Search?

Accepted Answer

Key features include vector similarity search, relevance scoring based on cosine similarity, complete transcript metadata (episode title, timestamps), configurable search parameters, and efficient database connection pooling.

Question 3

How do I configure Embedding Search?

Accepted Answer

Configuration is done through your MCP client (Cline or Claude Desktop) by adding the Embedding Search server details, including the path to the server's JavaScript file and environment variables for your Turso database URL and authentication token.

Question 4

What is Embedding Search?

Accepted Answer

Embedding Search is a Model Context Protocol (MCP) server that allows you to search transcript segments by leveraging vector similarity. It finds relevant segments based on a given query without creating new embeddings.

Question 5

What database schema is required for Embedding Search?

Accepted Answer

The tool expects a Turso database with two tables: `embeddings` (containing the vector embeddings and a foreign key to the transcripts table) and `transcripts` (containing the transcript text and metadata like episode title, start time, and end time).

Embedding Search

Embedding Search

主要功能

主要功能