A purpose, not a selector
Write what matters in plain English. TinyMonitor evaluates every change against it and marks the check meaningful or not, with a one-line explanation of what moved and why it passed.
/monitor
Subscribe to the web.
Hear only what matters.
TinyMonitor checks a page or a topic on a schedule, judges every change against a purpose you wrote, and delivers only the ones that meet it.
“Tell me when the Aurora 14 drops below $1,200 at any of our retail partners.”
TinyMonitor · Page · northgate.com/aurora-14 · daily 9:00
- Check 1Baseline captured
- Check 2Changed · not meaningful"Free shipping" banner reworded
- Check 3Changed · not meaningfulReview count 412 → 418
- Check 4Changed · meaningfulPrice $1,299 → $1,179
Northgate dropped to $1,179 this morning. First alert in four days. Nothing else was worth your attention.
Trusted by
Why TinyMonitor
Page or topic
Know the URL, watch the page: fetched fresh on every check, diffed against the last snapshot. Don’t know it yet, watch the topic: a search rerun on a recency window, flagging results that weren’t there before.
Judgment included
The change goes to our judge with your purpose. Your agent never reads the diff and never runs a model to decide if it mattered. The cost of the decision is ours.
Never silent on failure
Baseline, no change, meaningful, not meaningful, couldn’t judge, failed. Six outcomes, each visible, so a broken monitor can never look like a quiet one.
Use cases
For anything your agent keeps re-checking.
01
Pricing and availability
A competitor's plan price moves or a sold-out item is back. Old and new value, once.
02
People and organizations
A name added or removed on leadership, team, or investor pages.
03
Regulation and policy
New official guidance on your subject, with the document link. Commentary skipped.
04
Hiring signals
New roles at the companies you named, and nobody else.
05
Docs, changelogs, and terms
A rate limit, endpoint, or clause changes. Your agent knows before your users do.
Use it anywhere
Use TinyMonitor wherever your agents work.
Create a monitor from the API, CLI, MCP, or dashboard. Results land by email or webhook, on every agent surface.
REST API
MCP
CLI
Claude Code
Codex
OpenClawREST APIMCPCLIClaude CodeCodexOpenClawREST APIMCPCLIClaude CodeCodexOpenClawREST APIMCPCLIClaude CodeCodexOpenClawREST APIMCPCLIClaude CodeCodexOpenClawREST APIMCPCLIClaude CodeCodexOpenClawHow a check works
Every scheduled check ends in exactly one of these. Only the ones marked alert leave the dashboard.
| Outcome | What it means | Alerts |
|---|---|---|
| Baseline captured | First check. Watching from now. | No |
| No change | Same as last time. | No |
| Changed, meaningful | Met your purpose. | Yes |
| Changed, not meaningful | Moved, but not what you asked about. | No |
| Changed, couldn't judge | Diff available, judgment failed. | Yes |
| Failed | Page or schedule error, named. | Yes |
- Baseline capturedNo · alert
- First check. Watching from now.
- No changeNo · alert
- Same as last time.
- Changed, meaningfulYes · alert
- Met your purpose.
- Changed, not meaningfulNo · alert
- Moved, but not what you asked about.
- Changed, couldn't judgeYes · alert
- Diff available, judgment failed.
- FailedYes · alert
- Page or schedule error, named.
FAQ
Questions, answered.
What is TinyMonitor, and is it a web monitoring API?
TinyMonitor is a web monitoring API for agent workflows. It checks a page or a search topic on a schedule, compares each result with the last one, judges the change against a plain-English purpose, and delivers meaningful changes by email or webhook.
What is the difference between a Page monitor and a Topic monitor?
A Page monitor watches one URL and diffs the page content on every check. A Topic monitor reruns a search query with a recency window and reports results that are new since the last check. Use Page when you know the URL, Topic when you don't.
What is a purpose?
A purpose is a plain-English description of the change you care about. "The Pro plan price drops below $500." "A new research role opens at Anthropic. Ignore every other company." TinyMonitor judges every change against it. Without a purpose, TinyMonitor reports every change.
How is TinyMonitor different from a change detection tool?
Change detection tells you a page changed. TinyMonitor tells you whether the change you named happened. The judgment runs on TinyFish, so your agent receives a verdict and a summary instead of a diff to interpret.
When should I use TinyMonitor instead of calling TinySearch or TinyFetch on a schedule?
Use TinyMonitor when you'd otherwise be running the same TinySearch or TinyFetch on a loop and asking a model what changed. TinyMonitor owns the schedule, the snapshots, the comparison, and the judgment, and your agent only sees the result.
How often can a monitor run?
Page monitors run on any cron schedule you set. Topic monitors run at most every 30 minutes.
Stop checking. Start subscribing.
Use TinyMonitor for the pages and topics your agents keep re-reading, and get told when the change you named actually happens.