Do I need to write a launch.json file to debug in Flame?

For supported JavaScript and TypeScript scripts, use Debug Script without writing a launch.json file. Other runtimes need a configured external DAP adapter. CPU and heap profiling on this page use Flame’s built-in Node debug engine.

What breakpoint types and inspection features does Flame support?

Flame supports gutter line breakpoints, conditional breakpoints, and logpoints. When paused, you can step through execution, inspect Local, Closure, and Global variable scopes, monitor live watches, and view Monaco inline execution values beside your code.

How does the integrated CPU profiler and flamegraph viewer work?

Record CPU execution profiles with one click. Flame renders an interactive, zoomable canvas flamegraph alongside a Top Functions table showing self-time and total-time bottlenecks down to the exact millisecond and line number.

How do AI coding agents assist with debugging in Flame?

When execution pauses at an uncaught exception or breakpoint, you can send the live call stack, local variables, and execution context directly to AI. The agent diagnoses the root cause, explains the error state, and proposes verified code fixes.

Can Flame AI optimize CPU bottlenecks found in the profiler?

Yes. You can send profiling hotspots from the Top Functions table directly to an AI agent. The agent analyzes the CPU profile samples, pinpoints algorithmic inefficiencies, and refactors the code for improved performance.

AI debugger & profiler

AI Debugger & Profiler: 1-Click breakpoints, flamegraphs, and AI triage.

Launch supported Node.js scripts from the editor, pause at breakpoints, and inspect variables. Record CPU profiles with the built-in Node engine, send a measured hotspot to AI, and rerun the workload to check the result. Other runtimes use configured DAP adapters.

Download Flame See the debug workflow

The problem

Trace a failure while its runtime state is still available.

A stack trace names the failure, but a breakpoint lets you inspect the inputs and state that caused it. Start a supported Node.js script, reproduce the failing request, and pause where the behavior diverges. Step through the call stack and inspect values before changing the implementation.

For a slow request, record a representative CPU profile in the built-in Node engine. Inspect the hot functions and source lines, then decide whether the bottleneck is in the code you intended to optimize.

Flame unifies interactive debugging, CPU flamegraph profiling, and AI agent intelligence into one streamlined workflow. Click Debug Script to start stepping immediately, inspect live variables inline, and let AI agents explain stack traces and optimize hot code paths.

Integrated debugger and profilerZero-config breakpoints, live watches, inline execution values, and CPU flamegraphs.

How Flame solves it

From code issue to verified performance fix in five steps.

01
One-click debug launch
Use Debug Script for supported JavaScript and TypeScript workflows. Other runtimes require an appropriate configured Debug Adapter Protocol adapter.
02
Set gutter breakpoints & logpoints
Click the editor gutter to toggle line breakpoints, configure conditional pause expressions, or insert logpoints that print values without stopping execution.
03
Inspect variables & inline values
Pause execution to inspect Local, Closure, and Global scopes, monitor dynamic watches in real time, and view live evaluated values inline in Monaco.
04
Step & navigate call stacks
Step Over, Step Into, and Step Out with intuitive shortcuts. Click across stack frames to navigate call stacks and trace asynchronous boundaries across files.
05
Record CPU profiles & optimize with AI
Capture CPU execution profiles, explore zoomable canvas flamegraphs, and hand execution hotspots directly to AI agents to refactor performance bottlenecks.

Debugging & profiling features

Everything you need to diagnose bugs and optimize execution speed.

Feature 1

Zero-config 1-click launch

Launch debugging directly from breadcrumbs and file actions without hand-authoring launch configuration JSON files or memorizing CLI flags.

Feature 2

Gutter breakpoints & logpoints

Set standard line breakpoints, conditional pause expressions, and non-breaking logpoints directly in the editor gutter.

Feature 3

Scoped variable inspection

Expand Local, Closure, and Global scopes, add dynamic watch expressions, and evaluate arbitrary statements in the debug REPL console.

Feature 4

Monaco inline execution values

See live variable values and execution states rendered directly beside your source lines while stepping through code.

Feature 5

Interactive CPU flamegraphs

Capture CPU profile traces with an interactive zoomable canvas flamegraph and a Top Functions table highlighting self and total execution time.

Feature 6

AI root-cause diagnosis & triage

Send paused execution frames, stack traces, and flamegraph hotspots directly to AI agents to pinpoint bugs and refactor slow algorithms.

Video walkthrough

Watch Flame debug and profile code in 1 click.

Official Flame demo

Debug & Profile JavaScript & TypeScript in 1 Click

Watch how Flame launches debugging from editor actions, steps through code, inspects scoped variables, and captures CPU flamegraphs with AI optimization.

Flame on YouTube

Workflow comparison

Why Flame's AI debugger and profiler replaces manual setup and print debugging.

Flame AI Debugger & Profiler

1-click launch with zero manual launch.json configuration
Visual gutter breakpoints, conditional breaks, and logpoints
Monaco inline variable overlays showing live execution values
Interactive CPU canvas flamegraph with millisecond-level timings
AI agents analyze pause state and refactor performance hotspots

Traditional Debugging & Profiling

Hours spent debugging complex launch configurations and flags
Developers resort to scattering console.log statements across files
External profilers require separate browser windows and trace dumps
Disjointed tooling with no connection to coding agents
Manual trial-and-error to understand deep asynchronous stack errors

Debug frontend DOM, console, and network in the browser Let AI agents test and verify fixes in parallel worktrees Inspect and debug backend REST and GraphQL API routes

FAQ

AI debugger & profiler questions

Debug faster with zero configuration.

Download Flame to pause on breakpoints, inspect variables, and optimize code with CPU flamegraphs and AI assistance.

Download Flame Explore features

Free to use on macOS, Windows, and Linux.