rag-project-query

RAG Project Query — Project Knowledge Graph

Query the project-level LightRAG knowledge graph for architecture decisions, conventions, dependencies, and project-specific context.

Query Modes

Mode	When to use	What it does
hybrid (default)	Most queries	Combines local entity relationships + global topic summaries
local	"What's related to X in this project?"	Traverses entity relationships from specific nodes
global	"What are the project conventions?"	Summarizes across all project documents
naive	Simple keyword lookup	Basic text search, fastest but least intelligent

How to query

Full query

node -e "
const BASE = 'http://YOUR_LIGHTRAG_HOST:YOUR_PROJECT_PORT';
const API_KEY = process.env.LIGHTRAG_API_KEY || 'YOUR_API_KEY';
(async () => {
  const res = await fetch(BASE + '/query', {
    method: 'POST',
    headers: {
      'Content-Type': 'application/json',
      'X-API-Key': API_KEY
    },
    body: JSON.stringify({
      query: 'QUERY_HERE',
      mode: 'hybrid'
    })
  });
  const data = await res.json();
  console.log(JSON.stringify(data, null, 2));
})();
"

Context-only query

node -e "
const BASE = 'http://YOUR_LIGHTRAG_HOST:YOUR_PROJECT_PORT';
const API_KEY = process.env.LIGHTRAG_API_KEY || 'YOUR_API_KEY';
(async () => {
  const res = await fetch(BASE + '/query', {
    method: 'POST',
    headers: {
      'Content-Type': 'application/json',
      'X-API-Key': API_KEY
    },
    body: JSON.stringify({
      query: 'QUERY_HERE',
      mode: 'hybrid',
      only_need_context: true
    })
  });
  const data = await res.json();
  console.log(typeof data === 'string' ? data : JSON.stringify(data, null, 2));
})();
"

Response format

When presenting results:

• Distinguish project memory from personal memory — prefix with "Project context:" when relevant
• Be concise — summarize, don't dump raw JSON
• Cite the source type (e.g., "from an architecture decision stored on 2026-03-20")
• Flag stale info (>30 days old)
• If no results, say so — don't fabricate