Use case

Run scrapers and browser automation on a schedule

Scheduled scraping needs a runtime with a real browser, somewhere to keep the results, and a log to read when a site changes.

The problem

Serverless functions are too small and too short for a headless browser, and a cron job on your laptop only runs while the laptop is on.

What NeverHalt gives you

• Cron and webhooks

  Run on a 5-field cron schedule or start a job with an HTTPS call.

• Chromium and Playwright included

  The base image ships with Chromium and the libraries Playwright needs.

• A disk that survives restarts

  Every agent gets its own /app/data folder. SQLite files, checkpoints and caches are still there after a crash, a redeploy or a restart.

• You see what it does

  Live logs, plus an activity log of AI calls, tool calls and errors with token counts and cost.

Cron schedule

execution_mode: cron
cron_schedule: "*/30 * * * *"   # every 30 minutes

Questions

Can I trigger a run by hand?

Yes, from the dashboard or with an HTTPS call using the agent webhook token.

Are there limits on how often it runs?

Limits depend on your plan: the number of agents and the memory per agent.