Question 1

How does the AI code debugger work?

Accepted Answer

Pick a language (or let the tool auto-detect) and paste your code. The tool runs deterministic, pattern-based static analysis using per-language rule sets — finding undefined variables, missing returns, off-by-one loops, = vs == bugs, == vs === in JavaScript, mutable default arguments in Python, var usage, leftover console.log, TODO/FIXME markers, empty catch blocks, magic numbers, infinite loops, unreachable code after return, missing break in switch, hardcoded credentials, division-by-zero literals, and mismatched brackets. Each issue has a line number, severity, explanation, and suggested fix. The optional LLM enhancement uses your own key to explain tricky issues — never touching UnQTools servers.

Question 2

What languages are supported?

Accepted Answer

Python, JavaScript, TypeScript, Java, C++, and Go. Each language has its own rule set tuned to its common pitfalls. JavaScript/TypeScript rules flag var usage, == vs ===, and missing semicolons. Python rules flag mutable default arguments (def f(x=[])) and bare except. Java/C++ rules flag missing semicolons and missing break in switch. Go rules flag unused-likely patterns. Mismatched-bracket detection runs for all languages.

Question 3

Are the findings always correct?

Accepted Answer

No — and the tool is honest about it. Pattern-based static analysis catches common issues reliably (bracket mismatches, var declarations, console.log leftovers, TODO markers, = vs ==, missing break in switch). Other findings (undefined variables, missing returns, unreachable code) are heuristic and may produce false positives or miss real bugs — always review each finding before changing your code. The tool clearly labels findings as 'deterministic' (rule-based, reliable) vs 'heuristic' (best-effort, review needed). Optional BYO-key LLM enhancement can refine explanations.

Question 4

Can the tool auto-fix issues?

Accepted Answer

Yes — for safe, deterministic fixes. The 'Auto-fix' button applies only safe corrections: converting `var` to `let`, inserting missing semicolons, replacing `==` with `===` (and `!=` with `!==`) in JavaScript/TypeScript, and adding `break` statements to switch cases that are missing them. Each applied fix is listed so you can review what changed. Heuristic findings (undefined variables, missing returns, etc.) are NOT auto-fixed — they need your judgment.

Question 5

What extra features does this tool have?

Accepted Answer

(1) 6 languages with per-language rule sets. (2) Auto-detect source language. (3) 15+ bug-detection rules (undefined vars, missing returns, off-by-one, == vs ===, var usage, console.log, TODO/FIXME, empty catch, magic numbers, infinite loops, unreachable code, missing break, hardcoded credentials, division by zero, mismatched brackets). (4) Severity classification (info / warning / error / critical). (5) Line numbers + suggested fixes per issue. (6) Auto-fix button for safe corrections. (7) Side-by-side diff view (before vs after). (8) Runtime-error root-cause guesser (paste an error message). (9) Cyclomatic complexity hint. (10) Code-smell detection (long functions, deep nesting). (11) Sample buggy snippets per language. (12) Stats per severity. (13) Local history (localStorage, last 20). (14) Shareable URL with code + language encoded. (15) Optional BYO-key LLM explanation (OpenAI/Anthropic). (16) Honesty disclaimer about heuristic findings.

Question 6

Is my code sent anywhere?

Accepted Answer

No. All static analysis, auto-fixes, complexity calculation, and history run locally in your browser — your code never leaves this device. The only network path is if you explicitly paste your own LLM API key and click 'Explain with LLM' — that request goes directly to your chosen LLM provider (OpenAI or Anthropic) and never touches UnQTools servers. Skip the LLM step for 100% offline use.

AI Code Debugger & Bug Finder

How to use