# imsu.io

> imsu.io ("is my server up?") is an uptime monitoring service. It checks websites and HTTP APIs, TCP ports and cron jobs on a schedule, records whether each was up, how long it took and why it failed, and alerts the owner by email, Slack or Telegram when one goes down and when it comes back up. The free plan covers up to 10 monitors, each checked at most every 5 minutes, with no credit card.

## For agents

AI agents can manage a user's monitors (list, read, create, change, pause, resume, delete) through an MCP server at `https://imsu.io/mcp`, which signs in with OAuth, or through a REST API at `https://imsu.io/api/v1` with an API key the user creates. Only the user can sign up, sign in, approve access and create API keys, in the browser.

- [Start here: imsu.io for AI agents](https://imsu.io/docs/agents.md): which way to connect, what only the user can do, and what to do about each error
- [MCP server](https://imsu.io/docs/mcp.md): setup for Claude, Claude Code, Cursor, VS Code and other clients, and the tools
- [Authentication](https://imsu.io/docs/authentication.md): OAuth 2.1 for MCP clients, request by request
- [REST API](https://imsu.io/docs/api.md): API keys and every `/api/v1` operation with examples
- [Everything in one file](https://imsu.io/llms-full.txt): this file followed by all the docs pages

## Getting started

- [Home page](https://imsu.io/): what imsu.io does, how it works and the free plan
- [Create a free account](https://imsu.io/register): an email address and a password; no credit card
- [Sign in](https://imsu.io/login)
- [Privacy](https://imsu.io/docs/privacy.md): what personal data imsu.io keeps, including for status pages, the API and MCP, and how long

## What a monitor checks

- **HTTP:** any public `http://` or `https://` URL, requested with GET or HEAD. Up means a response within the timeout (1–30 seconds, 10 by default) with an expected status code (codes or ranges such as `200-299,301`; `200-299` by default) and, optionally, a keyword that must appear in the response body (GET only). Redirects are followed by default and can be turned off.
- **SSL certificates:** for `https://` URLs, an invalid certificate or failed TLS handshake marks the monitor down, labeled "SSL error". Alerts also go out before the certificate expires, at 30, 14, 7, 3 and 1 days left up to a chosen threshold (14 days by default; on by default for https). The issuer, subject and expiry date are shown per monitor and returned by the API and MCP ([SSL certificate checks](https://imsu.io/docs/api#ssl-certificate-checks)).
- **TCP port:** a connection to a host and port, e.g. a mail or database server. A reverse check alerts when a port that should stay closed accepts connections.
- **Heartbeat:** a cron job, backup or worker requests its own ping URL after each run (or the URL with `/fail` to report a failure). The monitor is down when no ping arrives within the expected interval plus a grace period. Intervals can be up to 31 days.
- Addresses that are not reachable from the public internet (private or local networks) cannot be monitored.
- Checks send the user agent `imsu.io-uptime-monitor/1.0` and come from the addresses listed at [https://imsu.io/api/meta/ips](https://imsu.io/api/meta/ips). [imsu.io bot](https://imsu.io/docs/bot.md) explains how to verify, allow or stop them.
- Interval: every 5 minutes at most on the free plan, up to once a day. Faster checks (down to every 10 seconds) can be enabled per account by imsu.io staff.
- No false alarms from a single dropped request: a failed check is retried within 30 seconds, and an HTTP or port monitor is marked down only after a set number of consecutive failures (1–10, 2 by default).
- Monitors can be paused and resumed.

## Alerts

- Sent when a monitor goes down (with the reason, e.g. a timeout, an unexpected status code or an SSL error) and when it is back up (with how long it was down), and before an https site's certificate expires.
- Channels: email (to the login address or another verified address), a Slack incoming webhook, and a Telegram bot. Each channel can be turned on and off separately.

## History and reports

- A dashboard with every monitor's current status, last response time and 24-hour uptime.
- Per monitor: individual checks with response times and errors, uptime and response time statistics, daily uptime bars, and every incident with its cause and duration.
- Individual check results are kept for 30 days; daily uptime figures for about 13 months.

## Plans

- Free: up to 10 monitors, checked every 5 minutes at most. No credit card. The MCP server and the REST API are included.
- Higher limits (more monitors, faster checks) are set per account by imsu.io staff; there is no self-service paid plan yet.

## Optional

- [AI catalog](https://imsu.io/.well-known/ai-catalog.json): machine-readable list of the resources imsu.io offers to AI agents
- [OpenAPI document](https://imsu.io/api/v1/openapi.json) of the REST API
- [Sitemap](https://imsu.io/sitemap.xml)