MergeCheck: Automation Result Verifier

Verify scheduled jobs and automations actually made the intended change, not just show green

Problem

Users rely on dashboard green checkmarks to judge automation success, but jobs run 'successfully' every day while the published content stays stale; existing monitoring only checks if the process crashed, not whether the output actually updated.

Solution

A lightweight monitoring service: after each automation run, verify via simple probes (file mtime, API response content, page text diff) that the expected change occurred; alert immediately if not, and record the actual change from each run.

Users

Indie developers, small teams, and ops people running scheduled content publishing, data sync, or ETL tasks.

MVP

Configure an HTTP probe or file check as the expected-change assertion; auto-run after each job and log results; alert via email/webhook on failure; simple dashboard showing true recent run status.

Revenue

Subscription by number of monitored tasks; free tier for 2 tasks; extra charges for alert volume or extended history retention.

Why now

AI agents and automation scripts are exploding, but tools to verify whether an agent actually completed its task are still scarce; indie developers can build fast with cron and simple probes.