- Published on
Moadim.io – A scheduler for agents
- Authors
- Name
- DeepHorus
Moadim, a new open-source loop engine for AI agents, has been launched to automate repetitive tasks across software repositories. The tool allows developers to define a scheduled loop comprising a prompt, a schedule, and a specific AI agent like Claude or Codex. This setup enables the agent to run against a code repository on a set schedule without manual intervention, marking a shift towards "loop engineering" over traditional prompting.
The software is positioned as a self-contained, local solution that operates without relying on cloud services or hidden queues. It uses a portable, in-process scheduler and runs each agent session inside an isolated tmux environment. This design ensures that each execution starts fresh, preventing state leakage between runs and allowing a watchdog process to terminate hung tasks.
Installation is straightforward, with options available via the Rust package manager Cargo or npm. The daemon, once installed, can be registered as a system service on macOS or Linux to ensure it persists across reboots. Notably, the tool requires no host cron daemon, as Moadim's internal scheduler manages all task timing.
Every loop created in Moadim is also exposed as a documented HTTP endpoint with an OpenAPI schema and a built-in Swagger UI. Furthermore, these loops function as tools within the Model Context Protocol (MCP), making them accessible to any compatible agent. This design choice simplifies integration and provides multiple interfaces for management and interaction.
The project is released under the MIT license, emphasizing its free, open-source nature for modification and self-hosting. Its FAQ confirms there is no paid tier or required account creation. Supported agents currently include Claude, Codex, Hermes, NanoClaw, and Pi, with the system providing isolation and management for each.
By introducing a robust scheduling and execution framework, Moadim aims to streamline the deployment of AI agents for routine codebase maintenance and analysis. The emphasis on local execution, isolation, and multiple access protocols addresses key concerns around reliability and security. As the tool gains traction, it could significantly alter how development teams integrate and manage automated AI-driven workflows.
