Scheduled monitoring with a disposable AI agent
A scheduled agent checks a page, a job log or a stock level every few minutes or every day and reports only when something changed. Ephemeral runs the saved workflow on your cron schedule and time zone, each time in a fresh microVM limited to the domains you allow, with the budget reserved up front and the verdict delivered by signed callback.
A page check took 19 seconds and cost $0.010047, measured on October 7, 2026 with openai/gpt-6-luna (2 model calls, 2,445 tokens).
What the run is allowed
Everything a run can touch is listed in its request. This is the request of the sample AI status page monitor: alert only when something changes.
| Files | None |
|---|---|
| Connections | None |
| Web access | ephemer.al, search off |
| Shell | Off |
| Budget | $0.03 per run, reserved before it starts |
| Timeout | 90 seconds |
How to run it
Start a run with the API or the CLI, or put it on a schedule with a cron expression and a time zone. The result arrives by signed callback or when you read the run.
POST /v1/runs
{
"workflow": "check-status-page",
"input": {
"url": "https://status.example.com"
},
"max_cost_micros": 30000
}Samples
Each sample is a complete request you can read, replay and copy.
- AI status page monitor: alert only when something changesOn a schedule, an agent reads a status page and returns one JSON verdict your own script can act on.
- Check that your scheduled jobs actually ran, with AIAn agent reads a list of cron jobs with their last run times and flags the overdue and failing ones in plain English.
- AI low-stock alert with reorder suggestionsAn agent reads inventory and eight weeks of sales, flags the SKUs that will run out soon and suggests what to reorder.
- Competitor pricing page change detector with AIAn agent fetches a pricing page, compares it with the snapshot you pass in, and explains what changed and why it matters.
- AI log analysis: group errors and find the likely root causeAn agent runs grep and sort over 200 lines of application logs inside a sandbox with no network, then names the likely cause and the first thing to check.
- Generate an incident postmortem draft with AIAn agent combines a status page incident with an affected payment record and fills a blameless postmortem template.
Recipe
Recipe: monitor a page walks through the same job step by step.
Related terms
- Run budget: The most a run may spend, reserved up front and enforced while it runs.
- MicroVM: The small virtual machine each run gets fresh and that is destroyed when the run ends.
Questions
How often can a check run?
A schedule takes a standard cron expression and a time zone. The status page sample runs every fifteen minutes, and you can preview the next run times before saving a schedule.
Can the agent visit any website?
Only the domains you list. Web access is off by default, and search is a separate switch, so a check can read the page you name and nothing else.
What stops a runaway schedule from costing a lot?
Every run has its own hard budget, reserved before the run starts, and a timeout. A run that reaches its budget stops there.
Prices are on the pricing page.