Skip to main content

Introduction to Cronicorn

GitHub Release npm version

Cronicorn is an HTTP job scheduler where endpoints read their own response data, adapt their timing, and trigger each other — controlled by plain English descriptions, not code. Set a baseline schedule, describe what matters, and let the AI handle the rest.

Getting Started​

PathLink
Web UI — create and manage jobs visuallycronicorn.com
MCP Server — manage jobs from Claude, Cursor, or any MCP clientMCP Server docs
API — integrate programmatically with the REST APIAPI Reference
Self-Host — run Cronicorn on your own infrastructureSelf-Hosting Guide

Quick Navigation​

Getting Started

Reference

Self-Hosting

Technical Deep Dive

What is Cronicorn?​

Cronicorn is a hosted scheduling service that replaces traditional cron with adaptive, AI-powered HTTP job scheduling. It is not a library, not an SDK, and there are no configuration files to deploy. You configure everything through the service — there is no code to write.

It combines:

  • Traditional cron scheduling - Set fixed intervals or cron expressions
  • AI-powered adaptation (optional) - Automatically adjust timing based on success rates, failure patterns, and response data
  • Plain English descriptions - Describe what matters in natural language, no config files or rules engines
  • HTTP endpoint execution - Call any HTTP endpoint on your schedule
  • Multi-endpoint coordination - Endpoints in the same job are aware of siblings and adapt together
  • Real-time monitoring - Track execution history, success rates, and performance

Why Use Cronicorn?​

Problems with Static Schedulers​

Traditional cron jobs have limitations:

  • Fixed timing - Run every 5 minutes whether needed or not
  • No learning - Repeated failures don't trigger automatic backoff
  • Manual intervention - You adjust schedules manually based on metrics
  • Resource waste - Over-polling wastes API rate limits and costs

How Cronicorn Helps​

For all users:

  • Set it and forget it - Define baseline schedules that just work
  • Real-time visibility - Track all executions with detailed history
  • Flexible scheduling - Use cron expressions or simple intervals
  • Constraint protection - Set min/max intervals to prevent over/under execution
  • Automatic error recovery - Exponential backoff and recovery actions out of the box

With AI enabled (optional):

  • Automatic adaptation - Schedules adjust based on actual response data
  • Intelligent backoff - Reduces frequency after failures automatically
  • Dynamic optimization - Increases frequency when needed, decreases when idle
  • Surge detection - Tightens polling during activity spikes, returns to baseline when stable
  • Always respects your constraints - AI suggestions stay within your min/max limits

Key Features​

🗓️ Flexible Scheduling​

  • Cron expressions: "0 */5 * * *" for traditional cron syntax
  • Simple intervals: 300000 (milliseconds) for straightforward timing
  • One-time runs: Schedule immediate or future single executions
  • Pause/resume: Temporarily disable endpoints for maintenance

🤖 AI Adaptation (Optional)​

When enabled, the AI planner:

  • Reads HTTP response bodies and interprets field values against your description
  • Suggests interval adjustments with expiration times (TTL)
  • Coordinates across sibling endpoints in the same job
  • Respects your configured min/max constraints
  • Gracefully degrades - baseline schedule continues if AI is unavailable

Note: Cronicorn works perfectly without AI. The baseline scheduler is production-ready and reliable. No per-endpoint AI setup is required — it runs automatically for all endpoints.

📊 Complete Visibility​

  • Run history: Every execution with timestamps, duration, status
  • Error tracking: Detailed error messages for failures
  • Success metrics: Track success rates and identify patterns
  • Scheduling transparency: See why each run happened (baseline, AI hint, etc.)

🛡️ Production Ready​

  • Reliable execution: Database-backed with distributed locking
  • Constraint protection: Min/max intervals prevent runaway schedules
  • Multi-tenant isolation: Secure separation between accounts
  • Three interfaces: Web UI, MCP Server, and HTTP API — all accept the same configuration
  • Self-hostable: Run on your own infrastructure with Docker Compose

How It Works​

  1. Create a Job - A container for related endpoints
  2. Add Endpoints - HTTP requests with baseline schedules and optional natural language descriptions
  3. Cronicorn executes them - The Scheduler worker makes HTTP calls on schedule
  4. AI adapts automatically (optional) - The AI Planner analyzes responses and adjusts frequency based on your descriptions
  5. Monitor results - View run history, AI decisions, and scheduling changes

You manage everything through the Web UI, MCP Server, or HTTP API — all three interfaces accept the same configuration.

Who Is Cronicorn For?​

  • SaaS developers monitoring API health across services
  • E-commerce teams syncing inventory and order data
  • DevOps engineers running scheduled maintenance tasks
  • Integration developers managing webhook retries and polling

Next Steps​