Cosa si dice in giro
Post, thread e citazioni su Hermes Agent raccolti da X/Twitter, GitHub Discussions e il Discord di Nous Research. Link all'originale sempre incluso.
"The agent that grows with you...." Hermes Agent README "The multi-instance release — Profiles for running isolated a..." v0.6.0 release notes "The agent that grows with you...." Hermes Agent README "The multi-instance release — Profiles for running isolated a..." v0.6.0 release notes Voci in evidenza
"The agent that grows with you."
"The self-improving AI agent built by Nous Research. It's the only agent with a built-in learning loop — it creates skills from experience, improves them during use, nudges itself to persist knowledge, searches its own past conversations, and builds a deepening model of who you are across sessions."
"The intelligence release — background task auto-notifications, free MiMo v2 Pro on Nous Portal, live model switching across all platforms, self-optimized GPT/Codex guidance, native Google AI Studio, smart inactivity timeouts, approval buttons, MCP OAuth 2.1, and 209 merged PRs."
"The multi-instance release — Profiles for running isolated agent instances, MCP server mode, Docker container, fallback provider chains, two new messaging platforms (Feishu/Lark and WeCom), Telegram webhook mode, Slack multi-workspace OAuth."
"Rolled Hermes out to replace three different chat bots on our team Slack. One config, one process, zero state drift. The official Docker container from v0.6.0 made the deploy boring, which is the highest praise I can give infrastructure."
"The self-optimized GPT/Codex tool-use guidance in v0.8.0 is quietly the most interesting PR in open-source agents this year. They are using benchmarks to regenerate prompts instead of hand-tuning them. That scales in a way hand-writing never will."
"Filesystem checkpoints with /rollback is the feature I show colleagues who are scared of letting agents touch their repos. It is the right answer to 'what happens if it does something bad' — the answer is 'you undo it.'"
"The resilience release — pluggable memory providers, credential pool rotation, Camofox anti-detection browser, inline diff previews, gateway hardening across race conditions and approval routing, and deep security fixes across 168 PRs."
"Runs on a $5 VPS, GPU clusters, or serverless infrastructure. Research-ready with batch trajectory generation and RL environments."
"My $5 Hetzner box runs Hermes as a Telegram bot, a Slack bot, and a cron-driven morning briefing. Load average is 0.03. Whatever they are doing to keep this lightweight, please do not stop."
"The ACP integration into VS Code is the thing I did not know I needed. I stopped context-switching to a terminal and never looked back. Feels like the agent lives in the editor now instead of hiding in a tab."
"Running Hermes on my old Pixel 4a via Termux and talking to it from Telegram on my current phone. Free, private, always on. It is deeply weird that a dead Android turned into my best self-hosted AI box."
"MCP OAuth 2.1 PKCE plus OSV malware scanning on extension packages is the first agent framework I have seen that treats plugin security like it matters. Adopted across my homelab the day v0.8.0 dropped."
"Same-provider credential pools in v0.7.0 saved me from a weekend of 401 firefights. The thread-safe least-used rotation just works with three Anthropic keys split across accounts. No more hand-picking which one is tired."
"I have tried every self-hosted AI assistant I could find in the last year. Hermes is the first one I did not uninstall after a week. Still here a month later. That has literally never happened before."
"The Honcho memory integration turned this from a chatbot into something that remembers how I work. It knows which projects I context-switch between, what my vocabulary is, when I am stuck. Not a feature — a different category of thing."
"Built a small Vue 3 project end-to-end with Hermes driving the edits from inside Zed via ACP. The inline diff previews added in v0.7.0 made the approval loop actually sane — I could see what it wanted to change before I agreed."
"Full MEDIA: tag delivery on Signal in v0.8.0 means I finally have a private, end-to-end encrypted line to my assistant. No more Telegram for anything sensitive. This is the upgrade I have been waiting on for months."
"Multi-instance Profiles in v0.6.0 let me run a work profile and a personal profile from one installation with zero credential leakage between them. Full token-lock isolation — the kind of boundary I used to rebuild myself badly."
"I introduced my non-technical partner to Hermes via Telegram for recipe help and now she uses it more than I do. Single best onboarding UX of any AI tool I have ever handed to a non-developer."
"Matrix tier-1 parity in v0.8.0 — reactions, read receipts, rich formatting, room management — made my self-hosted Synapse server feel like a first-class Hermes front door. I am finally off Telegram for good."
Hai scritto di Hermes?
Tagga @NousResearch su X o scrivi nel Discord di Nous Research e potremmo aggiungerti qui.