Cronicorn

HTTP jobs that understand their own responses

Features

• Adaptive scheduling — AI adjusts polling frequency based on real response data

• Plain English descriptions — no config files, no SDK required

• Safety constraints — min/max intervals and TTL-based hints that auto-expire

• Multi-endpoint coordination — endpoints are aware of siblings and adapt together

• Automatic error recovery — backoff, retry, and recovery actions out of the box

• Works without AI — traditional cron and interval scheduling as a fallback

Related MCP server: agentctl

How It Works

1. Add an HTTP endpoint and set a baseline schedule (cron or interval)

2. Describe what matters in plain English — thresholds, coordination logic, response structure

3. AI monitors responses and adapts frequency, triggers actions, and recovers from failures automatically

Getting Started