nrtsearch/searchLABSCO SUMMARY
Lucene syntax survives the trip through the assistant.
It does not paraphrase your query into something simpler — boolean groups, phrases, ranges and fuzzies go through as written, which matters when the index is the source of truth and the query has already been tuned against it.
What it is
A search tool over an NRTSearch/Lucene instance. It takes a full Lucene query — boolean, phrase, range, wildcard, fuzzy — against an index you name, and returns scored hits.
What you get
- Any Lucene query run against a named index
- The number of hits you want back, set per call
- The fields returned, chosen per call
- Matches highlighted on request
- Structured logging and retries around the call
Requirements
Python and a reachable NRTSearch instance. Point the server at its host and port through a config file, whose location an environment variable can override; the included quickstart installs dependencies and starts it.
Setup effort
Build from source — git clone https://github.com/tvergilio/nrtsearch-mcp-server.git
USE CASES
01
Searching an index without leaving the editor
The query runs from Copilot Chat or a desktop client, against the index you name in the call.
02
Running a query you already tuned
Boolean groups, phrase matches and range clauses are passed through intact rather than reinterpreted.
03
Getting back only the fields you need
Choose the retrieved fields per call, so a wide document does not arrive whole every time.
04
Seeing why a document matched
Highlighting shows the matching span, which is the difference between trusting a hit and checking it.
1 tool
Listed in project docs. Your client may show a different list.
BEFORE YOU INSTALL
What to know first
Tools you get
1 tool named in the project's documentation.
INSTALLATION
Connect your client
This one is built from source rather than installed with a single command.
Server setup
TERMINAL
git clone https://github.com/tvergilio/nrtsearch-mcp-server.git
cd nrtsearch-mcp-server
./quickstart.sh
