Persistproc
Enables large language models to manage and inspect long-running development processes.
About
Persistproc provides a shared process layer, acting as an MCP server and command-line tool, designed to streamline multi-agent development workflows. It allows AI agents and human developers to observe and control persistent processes like web servers, reducing the need for manual copy-pasting of errors or logs. By managing processes at runtime without a configuration file, Persistproc facilitates easier collaboration with multiple agents and remains tool-agnostic, enhancing developer productivity and interaction with AI-powered coding assistants.
Key Features
- Exposes a standard Model Context Protocol (MCP) server for programmatic interaction
- 3 GitHub stars
- Unified process control to start, stop, or restart managed processes
- Retrieves captured standard output and error streams from processes
- Ability to list all managed processes with filtering options and log paths
- Operates without a configuration file, managing processes entirely at runtime
Use Cases
- Allowing LLM agents to monitor and restart a single development web server, instantly reading its output.
- Enabling LLM agents to simultaneously read logs and diagnose issues from multiple local development services (e.g., API, frontend, SCSS builder, database).
- Integrating AI agents into existing development workflows to automatically manage and troubleshoot long-running processes.