Show HN: AgentX – Run Claude Code, Codex or Any LLM from Your Chats and GitHub

Category: ai-ml

Tags: ai-agents, self-hosted, multi-agent, automation, llm

Score: 7.0/10 (Innovation: 7, Technical: 8, Documentation: 8, Utility: 5)

AgentX is a self-hosted harness for running a team of AI agents on your own infrastructure, connecting them to chat channels, Git, schedules, and webhooks. It supports multiple model backends including Claude Code, Codex, and generic APIs, with a browser dashboard for monitoring. The project is ambitious and technically complex, but still experimental with significant setup friction and a narrow early-adopter audience.

Target audience: backend developers, devops, small engineering teams, AI automation enthusiasts

Repository: https://github.com/anis-marrouchi/agentx · TypeScript · MIT · 14 stars

View on Hacker News