Reproduce the bug.
Your agent fixes it.

Browse your app, hit a bug, drop a marker. Krometrail captures the full runtime context — browser state, network, console, framework state — so your coding agent can investigate and fix it without guessing.

Get Started View on GitHub

Reproduce

Use your app in Chrome like normal. Click around, fill forms, trigger the bug.

Mark

Hit the marker button when something goes wrong. Krometrail captures everything.

Hand off

Your agent gets the full session — network, errors, DOM, framework state — and debugs the code.

What Your Agent Sees

While you browse, Krometrail records everything via Chrome DevTools Protocol. When you drop a marker, your agent gets the full picture — no copy-pasting errors into chat.

Network requests

Every XHR/fetch with status, timing, headers, and response bodies — failed requests highlighted.

Console & errors

All console output and unhandled errors, timestamped and correlated with network events.

Framework state

React and Vue component trees, props, and state captured at each marker.

DOM & storage

DOM mutations, user interactions, localStorage/sessionStorage/cookie changes — all diffed.

Screenshots

Automatic screenshots at markers and on errors. Visual proof of what you saw.

Timeline

Everything timestamped and ordered. Your agent can search, inspect, and diff any moment.

Agent's view of your session

⚑ Marker: "checkout broke" at 00:51

Network
  POST /api/orders → 500 Internal Server Error (2.1s)
  Response: {"error": "discount_code_invalid"}

Console
  Error: Unhandled promise rejection: discount_code_invalid

Framework
  React: OrderForm state { code: "SAVE15", applying: true }
  → discountApplied never set (component stuck in loading state)

Screenshot
  [Captured at marker — spinner visible, no error shown to user]

Real Debugging, Not Guessing

Your agent sets breakpoints, steps through code, and inspects variables — just like you would. No more reading stack traces and hoping for the best.

①

②

③

④

── STOPPED at app/services/order.py:147 ──

Call Stack (3 frames)
  → order.py:147  apply_discount
    cart.py:89    checkout
    views.py:34   post

Source
  145│   subtotal = sum(item.price for item in items)
  146│   discount = calculate_discount(code)
→ 147│   final = subtotal * discount  # Bug: discount is -0.15
  148│   return Order(total=final)

Locals
  items     = [Item("Widget", 29.99), Item("Gadget", 49.99)]
  subtotal  = 79.98
  discount  = -0.15  ← expected 0.85
  code      = "SAVE15"

Language Support

10 languages via the Debug Adapter Protocol — one interface, any runtime.

PythonStable

debugpy

Node.js / TypeScriptStable

js-debug

GoStable

Delve

RustStable

GDB / LLDB

JavaStable

JDWP

C / C++Stable

GDB / lldb-dap

RubyStable

rdbg

C#Stable

netcoredbg

SwiftStable

lldb

KotlinStable

JDWP

Quick Start

Set up in 2 minutes. Works with any MCP-compatible agent.

json

{
  "mcpServers": {
    "krometrail": {
      "command": "krometrail",
      "args": ["--mcp"]
    }
  }
}

How Krometrail Compares

How Krometrail stacks up against other debugging tools for AI agents.

Feature	Krometrail	AIDB	mcp-debugger	mcp-dap-server
Viewport abstraction	✓	◐	✗	✗
Context compression	✓	✗	✗	✗
Browser observation	✓	✗	✗	✗
10+ languages	✓	◐	◐	◐
Conditional breakpoints	✓	✓	◐	✓
Watch expressions	✓	✗	✗	✗
Framework detection	✓	✓	✗	✗
CLI parity	✓	✗	✗	✗
Token awareness	✓	✗	✗	✗

✓ Full support ◐ Partial ✗ Not supported

Reproduce the bug.Your agent fixes it.