Cronloop

Cronloop is a hosted platform built by Cactal, Inc. that schedules and runs AI agents continuously on your behalf. You describe the job in plain Markdown, choose a model (Codex or Claude Code), set a cadence, and Cronloop handles execution, live output streaming, run history, and persistent memory across runs. It is available now with a free tier and a paid Pro plan.

What It Is

Cronloop sits in the category of autonomous agent scheduling and execution infrastructure. Rather than building a one-shot AI workflow, users define recurring agents — for sales outreach, support triage, SEO content updates, pull request generation, social posting, or competitor tracking — and Cronloop runs them on a cron-like schedule. Each run executes in a fresh sandbox, and each agent carries durable memory from one run to the next so context accumulates over time.

How the Agent Loop Works

The core workflow has three steps: write instructions in Markdown, pick a model provider connection (your existing ChatGPT or Claude subscription, or an Anthropic/OpenAI API key), and set a schedule. From there:

• Agents run as frequently as every 5 minutes (Pro) or every hour (Free)
• Each run streams live output to a dashboard
• Agents can install whatever tools the job needs inside a fresh sandbox per run
• Memory persists between runs, so agents build context over time
• Runs can be paused, stopped, or triggered on demand

Tool and Integration Surface

Cronloop connects to a wide range of third-party services via API, MCP, or CLI. The integration catalog spans:

• CRM and sales: HubSpot, Apollo, Salesforce
• Email and communication: Gmail, Intercom, Zendesk, WhatsApp, Telegram, Discord
• Marketing and SEO: Mailchimp, Google Ads, Google Analytics, Semrush, Google Search Console
• CMS and content: Webflow, WordPress, Shopify
• Finance: Stripe, QuickBooks, Xero, Wise
• Project management: Notion, Linear, Jira, Confluence, Asana, ClickUp, Trello, Airtable
• Dev and infra: GitHub, GitLab, Vercel, Netlify, Cloudflare, Docker, Kubernetes, PostgreSQL, MongoDB, Redis, Supabase, Firebase, Google Cloud, Snowflake, Datadog, Grafana, Sentry, PagerDuty

MCP Server and Assistant Integration

Cronloop exposes a hosted MCP server at https://app.cronloop.ai/v1/mcp, allowing users to manage agents directly from ChatGPT, Claude, or any MCP-compatible client. After a one-time OAuth sign-in, users can create agents, start runs, review run history, manage secrets, and read or write agent memory — all through natural language in their existing assistant. The MCP surface covers tools for agents, runs, configuration, environment variables, secrets, and memory.

Audience and Use Cases

The platform targets teams that want recurring AI automation without managing infrastructure. The homepage describes use cases including booking outbound demos, resolving support tickets automatically, growing organic search traffic, opening pull requests from issue trackers, automating social content, and tracking competitor pricing and launches. Agents run with the user's own model provider credentials, so model costs stay with the chosen provider rather than being bundled into Cronloop's subscription.