Give AI agents the power to schedule tasks, set reminders, and automate workflows. No cron syntax. No infrastructure. Just tasks that run on time.
Connect via MCP and your AI agents can create, manage, and monitor scheduled tasks autonomously. Results flow back via callbacks, enabling self-maintaining systems.
"every 15 minutes", "weekdays at 9am", "last friday of the month"
Task completes → agent receives results → agent reacts → creates new tasks
Critical operations require human approval before execution
Attach metadata to tasks—intent, conversation ID—returned in every callback
Give agents and product teams scheduling primitives that go beyond generic cron jobs.
Detect failures, trigger recovery steps, and keep callback context flowing back to the agent.
every 5 minutes
Let agents draft content, route it through review, and publish on a recurring cadence with callbacks.
daily at 9am
Increase frequency when health degrades, relax cadence after recovery, and keep run history visible.
every 30 minutes
Schedule user nudges, reports, and retries on behalf of each account without building custom schedulers.
weekdays at 5pm
Use the real SDK primitives that ship with Cronlet Cloud: tool definitions, tool execution handlers, callbacks, and task summaries.
import { getOpenAITools } from "@cronlet/sdk"
const response = await openai.chat.completions.create({
model: "gpt-5.2",
messages,
tools: getOpenAITools()
})import { getAnthropicTools } from "@cronlet/sdk"
const response = await anthropic.messages.create({
model: "claude-sonnet-4.6",
messages,
tools: getAnthropicTools()
})import { getLangChainTools } from "@cronlet/sdk"
const agent = createToolCallingAgent({
llm,
tools: getLangChainTools(client),
prompt
})npm install @cronlet/sdk
No credit card required. Upgrade when you need more runs or longer retention.
Perfect for prototyping and personal projects
For production workloads — $96/year (save 11%)
Enterprise-grade infrastructure without the enterprise complexity.
Linear or exponential backoff. Set max attempts, initial delays, and let us handle transient failures.
Per-task timeout limits prevent runaway jobs. Tasks are terminated cleanly when limits are exceeded.
Email or webhook notifications when tasks fail. Never miss a broken job again.
Store API tokens and credentials securely. Reference them in your tasks without exposing values.
Receive structured completion events after every run so agents can react, patch schedules, or create follow-up tasks.
Role-based permissions: owner, admin, member, viewer. Manage who can create, edit, or view tasks.
Need higher limits or custom requirements? Get in touch
No credit card required. 2,500 free runs every month.