We launched here in April as a way to run one private Hermes agent on Telegram without touching Docker or Fly. That worked — and then everyone who used it wanted a second agent, and then wanted to know what either one had actually done. So we rebuilt it. DeployHermes is now mission control for a squad: - A shared board — missions get delegated, claimed, and moved. You watch it instead of scrolling a chat log. - Agents with roles — each with its own model, memory, and skills. They hand work to each other. - Tickets — when an agent needs an approval or a judgment call, it stops and asks. Side-effecting actions go through that gate, not around it. - Runs — every execution is a receipt: what ran, what it cost, what it touched, why it failed. - 10 integrations live today — GitHub, Slack, Discord, Vercel, Google Workspace, Search Console, Sentry, DataFast, webhooks, and your own MCP servers. Telegram is still the fastest way to steer it. - MCP — Use the MCP to connect your Claude Code

Quick update on DeployHermes (managed hosting for Hermes agents on Fly.io). Since our last public release (right after we moved the stack to Vercel), we’ve been heads-down on reliability and on features people actually asked for. Here’s what’s new: Big additions Terminal access — including an non-interactive terminal path (with fixes and better behavior for sessions). Slack support — alongside improvements to Telegram and Discord (including clearer links for chat channels in the UI). Agent memory — memory on the agent side, plus home + settings UI so it’s usable, not buried. Cron jobs — create and manage scheduled work from the product (with follow-up fixes so jobs are recognized reliably). Skills — surface and use skills from the dashboard, with UI polish so detection and editing feel sane. Sessions & env — session management, an env editor, and a config editor so you’re not SSH’ing for every tweak. Multi-provider & OpenCode — multiple LLM providers