Sentry
Sentry is an error tracking and application monitoring tool: its SDK catches errors and slow requests in your web, backend or mobile app and groups them into issues, with the stack trace, logs, traces and a replay of the user's session attached.
It suits founders with a live app who want to hear about a bug before a customer writes in, and its MCP server lets Cursor and Claude Code read the issue when it is time to fix it.
The Developer plan is free for one user with 5,000 errors a month; Team is $29 a month, or $26 a month billed yearly, with 50,000 errors and pay-as-you-go above that, as checked on 30 October 2026.
Reviewed by the IndieHacked team
- What it does
- Error tracking for web, backend and mobile apps: what broke, for which user and in which release, with the trace and replay that led to it.
- Our take
- Founders with a live app who want to hear about a bug before a customer writes in. Add the SDK with one command, and when it is time to fix, let Cursor or Claude Code read the issue and its trace through Sentry's MCP server.
- Pricing
- FreemiumMid priceFree: the Developer plan, for one user, with 5,000 errors a month, email alerts and 30 days of history. Team is $29 a month, or $26 a month billed yearly ($312), and Business $89, or $80 a month billed yearly ($960), each with unlimited users, 50,000 errors, 5 GB of logs, 5 million spans, 50 replays and one uptime and one cron monitor a month. Above that quota you pay as you go: on Team from $0.0003625 an error, $0.50 a GB of logs, $1 an uptime monitor and $0.78 a cron monitor, up to a monthly budget you set. Without a budget nothing is charged beyond the plan, and data past the quota and budget is dropped for the rest of the month. Seer, the AI debugger, is $40 a month for each person who opens 2 or more pull requests in a connected repository. New accounts start with a 14-day Business trial, as checked on 30 October 2026.
The IndieHacked breakdown
What it does
Sentry's SDK goes into your app's code, with no agent to install on a server, and catches errors and slow operations as they happen. It groups errors into issues with a readable stack trace and links each one, by a shared trace ID, to the logs, spans, profiles and session replay around it, so you see the request or the screen that led to the failure. SDKs cover JavaScript, Next.js, React Native, Python, iOS, Android, Flutter and more, and the npx @sentry/wizard command sets a project up.
It also watches uptime and cron jobs and maps each issue to its release and suspect commit. Seer, the paid AI debugger, looks for the root cause, can open a pull request with a fix and reviews pull requests on GitHub or GitLab. The MCP server lets Cursor, Claude Code and other agents query issues and traces in plain language. Sentry's source is public, and a self-hosted version runs on your own server for free, with no dedicated support.
Who it suits
Founders with a web or mobile app in customers' hands. Errors arrive grouped, with the release, the commit and the user's session attached, so a fix starts from what happened rather than from a support email. One person can run it free on the Developer plan.
It fits founders who fix bugs with an AI agent. Sentry's own guides connect its MCP server to Cursor and Claude Code, so the agent reads the issue, the trace and the logs before it writes the fix, and Seer can hand its analysis to an external coding agent. Cursor and GitHub are both reviewed here.
Not the right fit for
Founders who want to see how people use the product: which features they open, where they drop out of a signup, who comes back. Sentry records sessions and metrics to explain errors and slow requests; funnels and retention are the job of a product analytics tool such as PostHog or Mixpanel, both reviewed here.
Developers who need to watch servers and infrastructure, not only the application. Sentry works from an SDK in your code with no agent on the host, and its Datadog integration puts Sentry's errors next to the metrics from other systems in Datadog's dashboards.
Report an error. Spotted a factual or language mistake? .