The moment your Lovable app ships to production, FlareLog catches every crash, silent API failure, and cost spike — automatically. Then brings the live logs into Lovable Agent so your AI can fix bugs without leaving the chat.
// Free up to 10,000 logs/mo · no credit card
100%
of Worker invocations seen
~3
lines of code to install
2 min
to first log in your dashboard
7–90 d
retention (free → Pro)
You don't have to wrap every function. The moment you add FlareLog, it watches your app end-to-end — frontend errors, API errors, and console logs — and sends them to your dashboard automatically.
Frontend crashes
React errors, broken buttons, blank screens — anything that breaks in the browser is captured with the stack trace.
API / backend errors
Failed login calls, checkout timeouts, database hiccups — FlareLog catches server errors without you writing a single log line.
AI debugging
Connect FlareLog to Cursor, Claude, Lovable, or your favorite AI with the MCP server. Ask "why did /checkout fail?" and get a real answer.
Structured logs
Call logger.info() anywhere to add extra context your AI can query later.
Every unhandled exception, console.error, and warning streamed instantly to your FlareLog dashboard with full stack traces. No more guessing.
Get emailed the second your app's backend usage starts spiking. No more surprise bills at the end of the month. Set a threshold and sleep easy.
Connect FlareLog's MCP server to Lovable, Cursor, or Claude Desktop. Your AI reads your production logs in-context and fixes bugs without leaving your editor.
One npm install and three lines of code. Works natively with the stack Lovable generates for you. No DevOps needed.
Full SDK support for TanStack Start server functions and API routes. Instrument your full-stack app with the same three lines.
Filter by level, worker, time range, or trace ID. Click any log to expand the full structured payload and stack trace. Designed for speed.
If you want session replay and deep JS profiling, Sentry is excellent. If you want error monitoring that works in 2 minutes and an AI that can read your live logs, FlareLog is for you.
01
Browser error trackers only see frontend crashes. FlareLog also captures server errors from your API routes, server actions, and database calls — the failures that actually break your app.
02
One npm install and three lines of code. No config files, no infrastructure, no manual wrapping. If you can paste a snippet into Lovable, you can use FlareLog.
03
The FlareLog MCP server and public skill repo let your AI read live logs and answer "why did /checkout fail?" without leaving the editor.
FlareLog's MCP server connects your production logs directly to Lovable Agent, Cursor, Claude Desktop, or Windsurf. Your AI can read real errors, filter by route, and suggest fixes — all without opening a browser tab.
Add MCP Server to Lovable Agent
Connect FlareLog to let Lovable Agent read live errors and traces directly in your prompt chat.
// Authentication Method
↑ Connectors → Add MCP server — no JSON, no terminal.
// cursor settings.json / claude_desktop_config.json
{
"mcpServers": {
"flarelog": {
"url": "https://mcp.flarelog.dev",
"headers": {
"Authorization": "Bearer fl_your_api_key"
}
}
}
}↑ Paste this into Cursor / Claude Desktop · Done.
Install the SDK
Run npm install @flarelog/sdk in your Lovable or TanStack Start project. It works with the stack Lovable already generates for you.
Add three lines of code
Import flarelog and call it with your API key. Console errors and unhandled exceptions are captured automatically — no wrapping needed.
Get your API key
Sign up free at flarelog.dev/login. Create a project, copy your key, done.
Connect the MCP server (optional)
In Lovable: open Connectors → Add MCP server, use https://mcp.flarelog.dev and paste your API key as the bearer token. In Cursor / Claude Desktop: paste the JSON config from the tab above.
// 1. Install the SDK
npm install @flarelog/sdk
import { flarelog } from "@flarelog/sdk";
import { FlareLogErrorBoundary, useFlareLog } from "@flarelog/sdk/react";
const logger = flarelog({ apiKey: process.env.VITE_FLARELOG_API_KEY, });
// Error Boundary
<FlareLogErrorBoundary logger={logger}>
<App />
</FlareLogErrorBoundary>
// src/start.ts
// add FLARELOG_API_KEY to your lovable secrets
import { createStart } from "@tanstack/react-start";
import { tanstackStartMiddleware } from "@flarelog/sdk/tanstack-start";
export const startInstance = createStart(() => ({
requestMiddleware: [tanstackStartMiddleware() as never],
}));
No — FlareLog and Sentry can run side-by-side. Many users keep Sentry for client-side session replay and use FlareLog for server errors and AI debugging. If you're a Lovable user just shipping your first app, FlareLog alone is usually enough.
The SDK adds sub-millisecond overhead per request. It batches and flushes logs asynchronously, so your users won't notice.
FlareLog captures failures at multiple layers, so even if something goes wrong before your code can log it, the error still shows up in your dashboard with the route, status, and any available context.
Yes. FlareLog itself has a free tier (10,000 logs/month) that's larger than most Lovable apps will ever need. No credit card, no DevOps experience required.
Yes — connect the MCP server to Lovable, Cursor, or Claude Desktop and ask things like "why did /checkout fail in the last hour?" or "show me every database timeout this week." The AI gets back the structured log payloads and can suggest code fixes in-context.
React, TanStack Start, and any setup Lovable exports. The React SDK ships an ErrorBoundary built for Lovable's generated stack. TanStack Start has a dedicated middleware. You can also use the raw SDK in custom server code.
No credit card, no DevOps experience required. Made for people who build with AI, not for people who configure Kubernetes.