Hermes Agent
Self-improving AI agent with multi-platform messaging gateway
https://git.due.ren/andreas/hermes-cloudron/-/raw/main/CloudronVersions.json
Description
Hermes Agent is a self-improving AI assistant by Nous Research. It features a built-in learning loop that creates and refines skills, searches past conversations, and builds user profiles across sessions.
Key Features:
- Multi-platform messaging: Telegram, Discord, Slack, WhatsApp, Signal, Matrix, Email, and more
- OpenAI-compatible API server for integration with Open WebUI, LobeChat, LibreChat, etc.
- Flexible LLM support: OpenRouter, OpenAI, Anthropic, xAI/Grok, and custom endpoints
- Procedural memory with autonomous skill creation
- Built-in cron scheduler for unattended tasks
- Parallel subagent processing
- Optional Nextcloud/WebDAV sync, Tailscale private networking, Bitwarden CLI, Cloudron CLI, local browser vision, xAI X search, Perplexity, and Vane search support in the Cloudron package
Recent Changes
Single-owner console capture + HUD lifecycle coverage.
Correct import order for window-renderer-lifecycle before window-reveal.
Wire renderer-lifecycle diagnostics into OAuth and portal login windows (#81290 follow-up).
Attach renderer-lifecycle diagnostics to all BrowserWindow instances.
Guard has_env profile probe with _safe like its sibling fields.
Raise fd soft limit + replace iterdir with scandir to stop fd leak.
Raise remote backend file limit.
Terminate app-managed gateway on shutdown.
Reap orphaned gateways before spawning restart.
Persist renderer crashes to desktop.log + finish the render() isolation class.
Package Hermes upstream main snapshot 0a60b164f56e.
