noodlbox

Context Search

Natural language code search

Search your codebase with natural language. Results are grouped by process and ranked by relevance.

Parameters

ParameterTypeRequiredDescription
repositorystringYesRepository name
qstringYesNatural language query
task_contextstringYesWhat you're working on
current_goalstringYesYour immediate objective
search_intentionstringYesWhy you're searching
include_contentbooleanNoInclude full symbol content (default: false)
limitnumberNoMax processes (1-100, default: 10)
max_symbolsnumberNoMax symbols per process (1-100, default: 20)
depthstringNoResult detail level: "definitions" (default) or "full"

Depth Parameter

Controls how much detail is returned for matched symbols:

ValueDescription
"definitions" (default)Symbol signatures and full body content only
"full"Symbols plus all relationships (calls, imports, edges)

Use "full" when you need to understand how code connects across the codebase.

Example

Ask your agent:

"Find code related to user authentication"

Or provide context for better results:

"I'm adding OAuth support. Find the existing authentication code so I can understand how to integrate with it."

Response

Results are grouped by process:

{
  "processes": [
    {
      "process_label": "UserAuthentication",
      "relevance_score": 0.95,
      "symbols": [
        {
          "symbol_name": "authenticateUser",
          "file_path": "src/auth/authenticate.ts",
          "line_number": 45
        }
      ]
    }
  ]
}

Tips

  • Be specific: "stripe webhook handler" beats "payment code"
  • Provide context: Better task descriptions = better results
  • Start broad, then narrow: Search "payments", then "refund processing"
  • Use include_content: true only when you need full function bodies

On this page