Cronicorn (weskerllc/cronicorn)

Cronicorn

https://github.com/weskerllc/cronicorn
Admin
Cronicorn is an AI job scheduler that dynamically adapts HTTP jobs to real-time system conditions,...

Tokens:124,507
Snippets:490
Trust Score:7.5
Update:2 months ago
Show doc for...
Context Summary (auto-generated)
Raw
# Cronicorn

Cronicorn is an intelligent HTTP scheduler that automatically adapts to your application's behavior. It replaces traditional cron jobs with AI-powered scheduling that adjusts execution frequency based on real response data, error patterns, and system load. Set baseline schedules, describe what matters in plain English, and let the AI optimize timing automatically while respecting your safety constraints.

The platform provides three interfaces for managing scheduled jobs: a Web UI at cronicorn.com, an MCP Server for AI assistants (Claude, Copilot, Cursor), and a REST API for programmatic access. Cronicorn works perfectly without AI as a reliable baseline scheduler, but when AI is enabled, it analyzes endpoint responses and adapts frequency—tightening monitoring during incidents and relaxing when stable. All AI suggestions have TTLs (time-to-live) and automatically revert to baseline, ensuring the system is self-healing.

## Authentication

Cronicorn supports two authentication methods: API keys for programmatic access and OAuth Device Flow for CLI tools and AI agents.

```bash
# API Key Authentication
# Generate keys in the web UI at /settings/api-keys
curl -H "x-api-key: cron_abc123..." \
  https://api.cronicorn.com/api/jobs

# OAuth Device Flow for CLI/AI tools
# Step 1: Request device code
curl -X POST https://api.cronicorn.com/api/auth/device/code
# Response: { "device_code": "DEVICE_CODE", "user_code": "ABCD-1234",
#             "verification_uri": "https://cronicorn.com/device", "expires_in": 1800 }

# Step 2: User authorizes in browser at verification_uri

# Step 3: Poll for token
curl -X POST https://api.cronicorn.com/api/auth/device/token \
  -H "Content-Type: application/json" \
  -d '{"device_code": "DEVICE_CODE"}'
# Response: { "access_token": "eyJ...", "token_type": "Bearer", "expires_in": 2592000 }

# Step 4: Use bearer token in requests
curl -H "Authorization: Bearer eyJ..." \
  https://api.cronicorn.com/api/jobs
```

## Create Job

Jobs are containers that group related endpoints together. Endpoints in the same job can coordinate through sibling visibility—the AI sees all their responses for multi-endpoint workflows.

```bash
curl -X POST https://api.cronicorn.com/api/jobs \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "API Health Checks",
    "description": "Monitor our production APIs with adaptive frequency"
  }'

# Response:
# {
#   "id": "job_abc123",
#   "name": "API Health Checks",
#   "description": "Monitor our production APIs with adaptive frequency",
#   "status": "active",
#   "createdAt": "2026-02-03T12:00:00Z"
# }
```

## List Jobs

Retrieve all jobs for your account with optional status filtering.

```bash
# List all active jobs
curl -H "x-api-key: YOUR_API_KEY" \
  "https://api.cronicorn.com/api/jobs?status=active"

# Get a specific job
curl -H "x-api-key: YOUR_API_KEY" \
  https://api.cronicorn.com/api/jobs/job_abc123

# Update a job
curl -X PATCH https://api.cronicorn.com/api/jobs/job_abc123 \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "name": "Production API Health Checks", "description": "Updated description" }'
```

## Job Lifecycle Management

Control job execution state with pause, resume, and archive operations.

```bash
# Pause a job (stops all endpoint executions)
curl -X POST https://api.cronicorn.com/api/jobs/job_abc123/pause \
  -H "x-api-key: YOUR_API_KEY"

# Resume a paused job
curl -X POST https://api.cronicorn.com/api/jobs/job_abc123/resume \
  -H "x-api-key: YOUR_API_KEY"

# Archive a job (permanent, removes from active list)
curl -X POST https://api.cronicorn.com/api/jobs/job_abc123/archive \
  -H "x-api-key: YOUR_API_KEY"
```

## Add Endpoint

Endpoints are HTTP requests that run on a schedule. Configure URL, method, baseline schedule (cron or interval), constraints, and a natural language description for AI adaptation.

```bash
curl -X POST https://api.cronicorn.com/api/jobs/job_abc123/endpoints \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Main API Health",
    "url": "https://api.example.com/health",
    "method": "GET",
    "baselineIntervalMs": 300000,
    "minIntervalMs": 30000,
    "maxIntervalMs": 900000,
    "timeoutMs": 30000,
    "description": "Monitors API health. Poll more frequently when status shows errors or error_rate_pct > 5%. Return to baseline when metrics normalize and error_rate_pct < 2%."
  }'

# Response:
# {
#   "id": "ep_xyz789",
#   "name": "Main API Health",
#   "jobId": "job_abc123",
#   "baselineIntervalMs": 300000,
#   "minIntervalMs": 30000,
#   "maxIntervalMs": 900000,
#   "createdAt": "2026-02-03T12:00:00Z"
# }
```

## Endpoint Configuration Fields

Complete reference for endpoint configuration options accepted by all interfaces (API, MCP Server, Web UI).

```json
{
  "name": "api-health-check",
  "url": "https://api.example.com/health",
  "method": "GET",
  "baselineIntervalMs": 300000,
  "minIntervalMs": 30000,
  "maxIntervalMs": 900000,
  "timeoutMs": 10000,
  "headersJson": { "Authorization": "Bearer token" },
  "bodyJson": null,
  "description": "Monitors API health. Poll more frequently when errors are detected."
}

// Field reference:
// name (required): What this endpoint does
// url (required): HTTP endpoint to call
// method (required): GET, POST, PUT, PATCH, or DELETE
// baselineIntervalMs: Milliseconds between runs (OR baselineCron)
// baselineCron: Cron expression e.g., "*/5 * * * *" (OR baselineIntervalMs)
// minIntervalMs: Minimum allowed interval (safety floor)
// maxIntervalMs: Maximum allowed interval (freshness guarantee)
// timeoutMs: Request timeout (default: 30000ms)
// headersJson: Custom HTTP headers object
// bodyJson: Request body for POST/PUT/PATCH
// description: Natural language instructions for AI adaptation
```

## List and Get Endpoints

Retrieve endpoints for a job or get details for a specific endpoint.

```bash
# List all endpoints in a job
curl -H "x-api-key: YOUR_API_KEY" \
  https://api.cronicorn.com/api/jobs/job_abc123/endpoints

# Get specific endpoint details
curl -H "x-api-key: YOUR_API_KEY" \
  https://api.cronicorn.com/api/jobs/job_abc123/endpoints/ep_xyz789

# Update an endpoint
curl -X PATCH https://api.cronicorn.com/api/jobs/job_abc123/endpoints/ep_xyz789 \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "baselineIntervalMs": 600000, "description": "Updated monitoring interval" }'
```

## Pause and Resume Endpoint

Temporarily pause an endpoint with an optional resume time.

```bash
# Pause until a specific time
curl -X POST https://api.cronicorn.com/api/endpoints/ep_xyz789/pause \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "pausedUntil": "2026-02-03T14:00:00Z",
    "reason": "Maintenance window"
  }'

# Resume immediately (set pausedUntil to null)
curl -X POST https://api.cronicorn.com/api/endpoints/ep_xyz789/pause \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "pausedUntil": null,
    "reason": "Maintenance complete"
  }'

# Archive an endpoint
curl -X POST https://api.cronicorn.com/api/jobs/job_abc123/endpoints/ep_xyz789/archive \
  -H "x-api-key: YOUR_API_KEY"
```

## Apply AI Interval Hint

Temporarily adjust execution frequency. Hints override the baseline schedule and expire automatically (TTL), reverting to baseline.

```bash
# Tighten monitoring to every 30 seconds for 1 hour
curl -X POST https://api.cronicorn.com/api/endpoints/ep_xyz789/hints/interval \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "intervalMs": 30000,
    "ttlMinutes": 60,
    "reason": "Increased monitoring during incident"
  }'

# Response:
# { "success": true, "intervalMs": 30000, "expiresAt": "2026-02-03T13:00:00Z" }
```

## Schedule One-Shot Execution

Trigger a specific one-time execution at a given time or immediately.

```bash
# Schedule immediate execution
curl -X POST https://api.cronicorn.com/api/endpoints/ep_xyz789/hints/oneshot \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "nextRunAt": "2026-02-03T12:30:00Z",
    "ttlMinutes": 30,
    "reason": "Immediate investigation after deployment"
  }'
```

## Clear Hints and Reset Failures

Return to baseline schedule and clear failure tracking.

```bash
# Clear all AI hints, revert to baseline schedule
curl -X DELETE https://api.cronicorn.com/api/endpoints/ep_xyz789/hints \
  -H "x-api-key: YOUR_API_KEY"

# Reset failure count (clears exponential backoff)
curl -X POST https://api.cronicorn.com/api/endpoints/ep_xyz789/reset-failures \
  -H "x-api-key: YOUR_API_KEY"
```

## List Endpoint Runs

Get execution history for an endpoint with optional filtering.

```bash
# Get recent runs with status filter
curl -H "x-api-key: YOUR_API_KEY" \
  "https://api.cronicorn.com/api/endpoints/ep_xyz789/runs?limit=20&status=failed"

# Response:
# {
#   "runs": [
#     {
#       "id": "run_abc123",
#       "endpointId": "ep_xyz789",
#       "status": "failed",
#       "statusCode": 503,
#       "durationMs": 1200,
#       "error": "Service Unavailable",
#       "startedAt": "2026-02-03T12:00:00Z"
#     }
#   ]
# }
```

## Get Run Details

Retrieve complete details for a specific execution run including response body.

```bash
curl -H "x-api-key: YOUR_API_KEY" \
  https://api.cronicorn.com/api/runs/run_abc123

# Response:
# {
#   "id": "run_abc123",
#   "endpointId": "ep_xyz789",
#   "status": "success",
#   "statusCode": 200,
#   "durationMs": 145,
#   "responseBody": { "healthy": true, "queue_depth": 45 },
#   "startedAt": "2026-02-03T12:00:00Z",
#   "completedAt": "2026-02-03T12:00:00.145Z",
#   "source": "baseline-interval"
# }
```

## Get Endpoint Health

Get aggregated health metrics for an endpoint over a time window.

```bash
curl -H "x-api-key: YOUR_API_KEY" \
  "https://api.cronicorn.com/api/endpoints/ep_xyz789/health?sinceHours=24"

# Response:
# {
#   "endpointId": "ep_xyz789",
#   "successCount": 285,
#   "failureCount": 3,
#   "successRate": 98.96,
#   "avgDurationMs": 142,
#   "lastRunAt": "2026-02-03T12:00:00Z",
#   "lastStatus": "success",
#   "failureStreak": 0
# }
```

## Get Dashboard Stats

Get aggregated statistics across all jobs and endpoints.

```bash
curl -H "x-api-key: YOUR_API_KEY" \
  "https://api.cronicorn.com/api/dashboard?startDate=2026-02-01&endDate=2026-02-03"
```

## AI Analysis Sessions

View AI decision history to understand why scheduling changes were made.

```bash
# List analysis sessions for an endpoint
curl -H "x-api-key: YOUR_API_KEY" \
  "https://api.cronicorn.com/api/endpoints/ep_xyz789/analysis-sessions?limit=10"

# Response:
# {
#   "sessions": [
#     {
#       "id": "session_abc123",
#       "endpointId": "ep_xyz789",
#       "createdAt": "2026-02-03T12:00:00Z",
#       "reasoning": "Queue depth increased from 50 to 200 over last 5 runs. Tightening monitoring interval from 5 minutes to 30 seconds.",
#       "toolsCalled": ["get_response_history", "propose_interval"],
#       "tokenUsage": 1250,
#       "durationMs": 3200
#     }
#   ],
#   "total": 45,
#   "hasMore": true
# }

# Get details for a specific session
curl -H "x-api-key: YOUR_API_KEY" \
  https://api.cronicorn.com/api/analysis-sessions/session_abc123
```

## JavaScript/TypeScript: Complete Health Monitoring Setup

Create a job with an adaptive health monitoring endpoint using the Fetch API.

```javascript
const API_KEY = 'YOUR_API_KEY';
const BASE_URL = 'https://api.cronicorn.com';

async function createHealthMonitoringJob() {
  // Step 1: Create the job
  const jobResponse = await fetch(`${BASE_URL}/api/jobs`, {
    method: 'POST',
    headers: {
      'x-api-key': API_KEY,
      'Content-Type': 'application/json'
    },
    body: JSON.stringify({
      name: 'Production API Monitoring',
      description: 'Monitors API health with adaptive frequency during incidents'
    })
  });

  if (!jobResponse.ok) {
    const error = await jobResponse.json();
    throw new Error(`Failed to create job: ${error.error?.message}`);
  }
  const job = await jobResponse.json();
  console.log('Created job:', job.id);

  // Step 2: Add endpoint with AI-adaptive description
  const endpointResponse = await fetch(`${BASE_URL}/api/jobs/${job.id}/endpoints`, {
    method: 'POST',
    headers: {
      'x-api-key': API_KEY,
      'Content-Type': 'application/json'
    },
    body: JSON.stringify({
      name: 'api-health-check',
      url: 'https://api.example.com/health',
      method: 'GET',
      baselineIntervalMs: 300000,    // 5 minutes
      minIntervalMs: 30000,          // 30 seconds floor
      maxIntervalMs: 900000,         // 15 minutes ceiling
      timeoutMs: 10000,
      description: 'Monitor API health. When error_rate_pct > 5%, tighten polling to 30 seconds. Return to baseline when error_rate_pct < 2% and status is healthy.'
    })
  });

  if (!endpointResponse.ok) {
    const error = await endpointResponse.json();
    throw new Error(`Failed to create endpoint: ${error.error?.message}`);
  }
  const endpoint = await endpointResponse.json();
  console.log('Created endpoint:', endpoint.id);

  return { job, endpoint };
}

// Helper functions for runtime management
async function tightenMonitoring(endpointId, intervalMs = 30000, ttlMinutes = 60) {
  const response = await fetch(`${BASE_URL}/api/endpoints/${endpointId}/hints/interval`, {
    method: 'POST',
    headers: { 'x-api-key': API_KEY, 'Content-Type': 'application/json' },
    body: JSON.stringify({ intervalMs, ttlMinutes, reason: `Manual override: ${intervalMs}ms for ${ttlMinutes} min` })
  });
  if (!response.ok) throw new Error(`Failed to apply hint: ${response.statusText}`);
}

async function returnToBaseline(endpointId) {
  await fetch(`${BASE_URL}/api/endpoints/${endpointId}/hints`, {
    method: 'DELETE',
    headers: { 'x-api-key': API_KEY }
  });
}

async function checkHealth(endpointId, sinceHours = 24) {
  const response = await fetch(`${BASE_URL}/api/endpoints/${endpointId}/health?sinceHours=${sinceHours}`, {
    headers: { 'x-api-key': API_KEY }
  });
  return response.json();
}

createHealthMonitoringJob().catch(console.error);
```

## Python: Create Job and Dynamic Scheduling

Python implementation for creating jobs and adjusting scheduling based on external metrics.

```python
import requests

API_KEY = "YOUR_API_KEY"
BASE_URL = "https://api.cronicorn.com"
HEADERS = {"x-api-key": API_KEY, "Content-Type": "application/json"}

# Create job and endpoint
job_response = requests.post(
    f"{BASE_URL}/api/jobs",
    headers=HEADERS,
    json={
        "name": "Production API Monitoring",
        "description": "Monitors API health with adaptive frequency"
    }
)
job_response.raise_for_status()
job = job_response.json()
print(f"Created job: {job['id']}")

endpoint_response = requests.post(
    f"{BASE_URL}/api/jobs/{job['id']}/endpoints",
    headers=HEADERS,
    json={
        "name": "api-health-check",
        "url": "https://api.example.com/health",
        "method": "GET",
        "baselineIntervalMs": 300000,
        "minIntervalMs": 30000,
        "maxIntervalMs": 900000,
        "timeoutMs": 10000,
        "description": "Poll every 30s when status is degraded or error_rate_pct > 5%. Return to baseline when healthy."
    }
)
endpoint_response.raise_for_status()
endpoint = endpoint_response.json()
print(f"Created endpoint: {endpoint['id']}")

# Dynamic scheduling based on external metrics
def adjust_scheduling_from_metrics(endpoint_id):
    """Check external metrics and apply Cronicorn scheduling hints."""
    metrics = requests.get("https://your-monitoring.com/api/metrics").json()
    error_rate = metrics.get("error_rate_pct", 0)
    cpu_load = metrics.get("cpu_load_pct", 0)

    if error_rate > 10 or cpu_load > 90:
        # Critical: tighten to 30 seconds
        requests.post(
            f"{BASE_URL}/api/endpoints/{endpoint_id}/hints/interval",
            headers=HEADERS,
            json={"intervalMs": 30000, "ttlMinutes": 30, "reason": f"Critical: error={error_rate}%, cpu={cpu_load}%"}
        )
        print(f"Tightened to 30s")
    elif error_rate > 5:
        # Warning: tighten to 1 minute
        requests.post(
            f"{BASE_URL}/api/endpoints/{endpoint_id}/hints/interval",
            headers=HEADERS,
            json={"intervalMs": 60000, "ttlMinutes": 15, "reason": f"Warning: error_rate={error_rate}%"}
        )
    else:
        # Normal: clear hints
        requests.delete(f"{BASE_URL}/api/endpoints/{endpoint_id}/hints", headers={"x-api-key": API_KEY})
        print("Normal: returned to baseline")
```

## MCP Server Setup

Install and configure the MCP Server to manage Cronicorn through AI assistants like Claude, Copilot, or Cursor.

```bash
# Install globally
npm install -g @cronicorn/mcp-server

# Claude Code CLI
claude mcp add cronicorn -- npx -y @cronicorn/mcp-server

# Claude Desktop / Copilot / Cursor (add to config file)
# {
#   "mcpServers": {
#     "cronicorn": {
#       "command": "npx",
#       "args": ["-y", "@cronicorn/mcp-server"]
#     }
#   }
# }

# First use triggers OAuth in your browser. Approve once, stay connected for 30 days.

# Example conversations:
# "Check my API health every 5 minutes"
# "Show me why that endpoint is failing"
# "Migrate my 10 Vercel cron jobs to Cronicorn"
```

## MCP Tool Call Examples

Concrete examples of MCP tool invocations showing inputs and expected responses.

```json
// createJob
// Input:
{ "name": "Production API Monitoring", "description": "Monitors API health with adaptive frequency" }
// Response:
{ "id": "job_abc123", "name": "Production API Monitoring", "status": "active" }

// addEndpoint
// Input:
{
  "jobId": "job_abc123",
  "name": "api-health-check",
  "url": "https://api.example.com/health",
  "method": "GET",
  "baselineIntervalMs": 300000,
  "minIntervalMs": 30000,
  "description": "Poll every 30s when status is degraded. Return to baseline when healthy."
}
// Response:
{ "id": "ep_xyz789", "name": "api-health-check", "jobId": "job_abc123" }

// applyIntervalHint
// Input:
{ "id": "ep_xyz789", "intervalMs": 30000, "ttlMinutes": 60, "reason": "Incident detected" }
// Response:
{ "success": true, "intervalMs": 30000, "expiresAt": "2026-02-03T13:00:00Z" }

// getEndpointHealth
// Input:
{ "id": "ep_xyz789", "sinceHours": 24 }
// Response:
{ "successCount": 285, "failureCount": 3, "successRate": 98.96, "failureStreak": 0 }

// listEndpointRuns
// Input:
{ "id": "ep_xyz789", "limit": 5, "status": "failed" }
// Response:
{ "runs": [
  { "id": "run_001", "status": "failed", "statusCode": 503, "error": "Service Unavailable" }
]}

// resetFailures
// Input:
{ "id": "ep_xyz789" }
// Response:
{ "success": true, "failureCount": 0 }
```

## Self-Hosting with Docker Compose

Run Cronicorn on your own infrastructure using Docker Compose.

```bash
# Download docker-compose.yml from GitHub
# Create .env file with required secrets

# Minimal .env for local use:
BETTER_AUTH_SECRET=your-random-32-character-secret-here

# Production .env:
BETTER_AUTH_SECRET=your-random-32-character-secret-here
WEB_URL=https://yourdomain.com
API_URL=http://cronicorn-api:3333
BETTER_AUTH_URL=https://yourdomain.com

# Optional: GitHub OAuth
GITHUB_CLIENT_ID=your_github_client_id
GITHUB_CLIENT_SECRET=your_github_client_secret

# Optional: AI Scheduling
OPENAI_API_KEY=sk-your-key-here
AI_MODEL=gpt-4o-mini

# Start services
docker compose up -d

# Access:
# - Web app: http://localhost:5173
# - API: http://localhost:3333
# - Default login: admin@example.com / devpassword

# Useful commands
docker compose logs -f          # View logs
docker compose logs -f api      # Specific service logs
docker compose restart          # Restart services
docker compose pull && docker compose up -d  # Update to latest
```

## Response Body Design for AI Adaptation

Structure your endpoint response bodies to include fields the AI can interpret. The AI reads response bodies automatically—no parsing code needed.

```json
// Health monitoring response
{
  "status": "degraded",
  "error_rate_pct": 8.5,
  "latency_ms": 1200,
  "healthy": false,
  "timestamp": "2026-02-03T12:05:00Z"
}

// Data sync response with volume metrics
{
  "records_pending": 15000,
  "sync_rate_per_minute": 500,
  "estimated_completion_minutes": 30,
  "status": "syncing"
}

// System load with inverse scaling signals
{
  "cpu_pct": 85,
  "memory_pct": 72,
  "load_avg_1m": 4.2,
  "recommendation": "reduce_polling"
}

// Stability-focused with smoothed values
{
  "instant_value": 523,
  "avg_5min": 487,
  "avg_1hr": 502,
  "trend": "stable",
  "within_normal_range": true
}

// Coordination signals for multi-endpoint workflows
{
  "status": "error",
  "error_count": 15,
  "needs_recovery": true,
  "last_error": "Connection refused"
}
```

## Exponential Backoff and HTTP Status Handling

Cronicorn automatically applies exponential backoff on failures and resets on success.

```bash
# Built-in behavior (automatic):
# - 2xx responses → Success, failure count resets to 0
# - 4xx/5xx responses → Failure, backoff applies (baseline × 2^failures, max 32×)
# - Timeouts → Treated as failure

# Backoff multipliers:
# Failures | Multiplier | 5-min baseline becomes
#    0     |    1×      | 5 minutes
#    1     |    2×      | 10 minutes
#    2     |    4×      | 20 minutes
#    3     |    8×      | 40 minutes
#    4     |   16×      | 80 minutes
#   5+     |   32× cap  | 160 minutes

# AI can override backoff during incidents:
curl -X POST https://api.cronicorn.com/api/endpoints/ep_xyz789/hints/interval \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "intervalMs": 30000,
    "ttlMinutes": 15,
    "reason": "Override backoff - actively monitoring failing endpoint"
  }'

# Manual reset to clear backoff
curl -X POST https://api.cronicorn.com/api/endpoints/ep_xyz789/reset-failures \
  -H "x-api-key: YOUR_API_KEY"
```

## Multi-Endpoint Recovery Workflow

Configure coordinated recovery automation where a health check endpoint triggers a recovery action endpoint.

```bash
# Create job for both endpoints (same job = sibling visibility)
curl -X POST https://api.cronicorn.com/api/jobs \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{ "name": "Service Recovery Automation" }'
# Returns: { "id": "job_abc123" }

# Add health-check endpoint
curl -X POST https://api.cronicorn.com/api/jobs/job_abc123/endpoints \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "health-check",
    "url": "https://api.example.com/health",
    "method": "GET",
    "baselineIntervalMs": 300000,
    "minIntervalMs": 30000,
    "description": "Monitors service health. When needs_recovery is true, the trigger-recovery sibling should run. Tighten to 30s during errors."
  }'

# Add recovery endpoint (sibling of health-check)
curl -X POST https://api.cronicorn.com/api/jobs/job_abc123/endpoints \
  -H "x-api-key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "name": "trigger-recovery",
    "url": "https://api.example.com/admin/restart",
    "method": "POST",
    "baselineIntervalMs": 86400000,
    "minIntervalMs": 300000,
    "description": "Recovery action. Only run when health-check sibling shows needs_recovery=true. Wait 5 minutes between attempts. Max 3 attempts before pausing 1 hour."
  }'

# AI behavior:
# 1. health-check returns {"needs_recovery": true}
# 2. AI calls get_sibling_latest_responses() to see trigger-recovery
# 3. AI calls propose_next_time() on trigger-recovery for immediate run
# 4. AI calls propose_interval(30000) on health-check for close monitoring
# 5. After recovery: AI calls clear_hints() on both → back to baseline
```

## Summary

Cronicorn is ideal for scenarios where static cron jobs fall short: monitoring systems that need adaptive frequency during incidents, data pipelines with volume-based polling, coordinated multi-endpoint workflows, and any situation where you adjust schedules manually based on metrics. The platform excels at flash sale monitoring (automatic tightening during traffic surges), ETL pipeline coordination (dependent endpoints wait for upstream success), infrastructure auto-remediation (investigating and fixing issues before paging engineers), and competitive price monitoring (respecting rate limits while intensifying during competitor sales).

Integration patterns include: using the REST API for programmatic control from CI/CD and scripts, the MCP Server for conversational management through AI assistants, and the Web UI for visual monitoring and configuration. All three interfaces share the same underlying data model and support identical configurations. For production deployments, Cronicorn can be self-hosted via Docker Compose with your own reverse proxy for HTTPS, or used as a hosted service at cronicorn.com with GitHub OAuth authentication. The system is designed to be self-healing—AI hints expire automatically, endpoints continue on baseline schedules if AI becomes unavailable, and a single successful response resets all backoff logic.