The Code Finder endpoints provide precise, targeted lookup into the knowledge graph. Unlike the unified search endpoint, each Code Finder route targets a specific node type (function, class, variable, module, import) or capability (file peeking, complexity analysis). Use these endpoints when you already know what kind of symbol you are looking for, or when you need to inspect a specific line range in a file stored in the graph.Documentation Index
Fetch the complete documentation index at: https://personal-ce79cb71.mintlify.app/llms.txt
Use this file to discover all available pages before exploring further.
GET /api/v1/query/code-finder/function
Finds function nodes by name. Supports exact and fuzzy matching.Function name to search for.
Enable fuzzy matching to find functions with similar names.
Filter to a specific repository owner.
Filter to a specific repository name.
Response
The query name.
Whether fuzzy matching was enabled.
Matching function nodes with name, path, line_number, and source_code.
Number of matches returned.
GET /api/v1/query/code-finder/class
Finds class nodes by name. Supports exact and fuzzy matching.Class name to search for.
Enable fuzzy matching.
Filter to a specific repository owner.
Filter to a specific repository name.
Response
The query name.
Whether fuzzy matching was enabled.
Matching class nodes.
Number of matches returned.
GET /api/v1/query/code-finder/variable
Finds variable nodes whose name contains the given substring.Variable name or substring to search for.
Filter to a specific repository owner.
Filter to a specific repository name.
GET /api/v1/query/code-finder/content
Finds functions, classes, and variables whose source code contains a given pattern. Useful when you remember a code pattern but not the symbol name.Pattern to match against the
source_code property of graph nodes.Filter to a specific repository owner.
Filter to a specific repository name.
GET /api/v1/query/code-finder/module
Finds module nodes by name. Module nodes represent files or packages as stored in the graph.Module name to search for.
GET /api/v1/query/code-finder/imports
Finds import statement nodes by import name or alias.Import name or alias to search for (e.g.
fastapi, APIRouter, np).Filter to a specific repository owner.
Filter to a specific repository name.
GET /api/v1/query/code-finder/peek
Returns a window of lines from a file stored in the graph, centered on an anchor line. Use this endpoint to view context around a result fromcode-finder/line or any endpoint that returns a path and line_number.
File path as stored in the graph. This is the absolute or repo-relative path recorded during ingestion.
Anchor line number (1-indexed). Defaults to
1 if omitted. The anchor line is flagged with is_anchor: true in the response.Number of lines to include above the anchor. Maximum 200.
Number of lines to include below the anchor. Maximum 200.
Filter to a specific repository owner.
Filter to a specific repository name.
Response
Array of line objects in the requested window.
Example
GET /api/v1/query/code-finder/complexity
Returns the cyclomatic complexity score and risk level for a specific function.Name of the function to analyze.
Optional file path filter. Use this when multiple functions share the same name across different files.
Filter to a specific repository owner.
Filter to a specific repository name.
Response
The queried function name.
The path filter applied, or
null.Complexity result including the numeric score and a risk level label.
Error responses
| Status | Cause |
|---|---|
404 | No function with that name (and optional path) was found in the graph. |
GET /api/v1/query/code-finder/complexity/top
Returns the top N most complex functions across the graph, ranked by cyclomatic complexity score in descending order. Use this to identify the highest-risk functions in a codebase.Number of functions to return.
Filter to a specific repository owner.
Filter to a specific repository name.
Response
The requested limit.
Functions ranked by complexity score, highest first. Each object includes name, path, line_number, complexity score, and risk level.
Number of results returned.