Why it matters
HotPlex aims to simplify the integration and management of AI coding agents across different platforms. By offering a unified access layer and robust features like session management, multi-platform delivery, and enhanced security, it could streamline development workflows for AI-powered coding assistance and agent deployment.

HotPlex is a high-performance Go gateway that acts as a unified bridge for AI coding agents. It provides a single WebSocket interface (AEP v1) for accessing various AI coding agents, including Claude Code, OpenCode Server, ACP, and Codex CLI. The gateway features backpressure-aware streaming, per-session monotonic sequencing, and automatic LLM retry with exponential backoff. It supports multi-platform delivery, allowing agents to be deployed on Web, Slack, and Feishu without code changes, and includes platform-native streaming, slash commands, and interaction management. HotPlex also offers deterministic sessions using UUIDv5 for seamless reconnections, with support for SQLite/PostgreSQL persistence and background garbage collection. A notable feature is its AI-native cron scheduler, enabling agents to create scheduled tasks from natural language. The project includes an embedded web chat and an admin UI for bot management, API keys, and session monitoring. Security features include timing-safe API key authentication, SSRF protection, command/tool/model allowlists, and environment isolation. For operations, it provides over 30 Prometheus metrics, OpenTelemetry distributed tracing, and structured JSON logging. HotPlex also comes with a self-contained CLI for various tasks, including gateway management, service control, guided setup, diagnostics, and configuration hot-reloading.

Share:XHacker NewsLink
Article ID - cmpzmkfja0Featured on AI Radar: HotPlex: A Unified Access Layer for AI Coding Agents