{"slug": "hermes-agent-on-nvidia-jetson-orin-nano-practical-setup-guide-config-and-health", "title": "Hermes Agent on NVIDIA Jetson Orin Nano — Practical setup guide, config, and health-check script for running a self-hosted AI agent on edge hardware", "summary": "A developer successfully deployed Hermes Agent on an NVIDIA Jetson Orin Nano (8 GB) for self-hosted AI agent runtime, achieving a silent, low-power automation box running 24/7. The setup uses Ubuntu 22.04 aarch64 with JetPack 6.2, system Chromium for browser automation, and power mode switching via nvpmodel to balance performance and thermals. Key observations include 5 GB usable RAM for agent and browser, NVMe-backed swap for memory pressure, and remote API for LLM inference due to limited VRAM.", "body_md": "Practical setup tips, performance observations, and gotchas from running a self-hosted AI agent runtime on a Jetson Orin Nano (8 GB) in production-like conditions.\n\nThe Jetson Orin Nano is an odd duck: ARM64, 7–8 GB RAM shared between CPU/GPU, no x86 emulation shortcuts, and a Linux kernel that's just different enough to trip you up. But it's also:\n\n**Cheap to run**(~15 W under load)** Silent**(fan-cooled, no coil whine)** Persistent**(NVMe boot, can run 24/7)** Surprisingly capable**for an edge device running a full agent stack\n\nIf you want a dedicated automation box that sits on a shelf and runs Hermes Agent around the clock, this is a solid choice.\n\n| Item | Value |\n|---|---|\nSoC |\nNVIDIA Orin Nano (8 GB unified memory) |\nCPU |\n6× Cortex-A78AE @ up to 1.5 GHz |\nRAM |\n7.4 GB (shared GPU/CPU via unified memory) |\nStorage |\nNVMe SSD (boot + agent data) |\nOS |\nUbuntu 22.04 aarch64 (JetPack 6.2 / R36) |\nKernel |\n5.15.185-tegra |\nAgent |\nHermes Agent v2026.5.16 |\nPower mode |\nMAXN_SUPER (default) |\n\nHermes Agent needs Chromium for browser automation and a few Python extras. The JetPack repos have everything:\n\n```\n# Browser runtime for Playwright-based tools\nsudo apt install chromium-browser\n\n# System deps for Playwright\nsudo apt install libgstreamer1.0-0 libgstreamer-plugins-base1.0-0 \\\n                 gstreamer1.0-plugins-good gstreamer1.0-plugins-bad \\\n                 libva-drm2\n\n# Hermes Agent's Python deps want numpy, so get the accelerated build\nsudo apt install python3-numpy\n```\n\n**Gotcha:** Don't install `python3-pip`\n\n's Chromium via `playwright install`\n\n— it downloads x86 binaries that won't run. Use the system `chromium-browser`\n\nand point Playwright at it with `PLAYWRIGHT_BROWSERS_PATH=/usr/bin chromium-browser`\n\n.\n\nThe Orin Nano has several power modes exposed via `nvpmodel`\n\n. The default is `MAXN_SUPER`\n\n(all 6 cores, max clocks), but you can trade performance for thermals:\n\n```\n# List available modes\nsudo nvpmodel -q\n\n# Modes typical on Orin Nano 8 GB:\n#   MAXN_SUPER  — 6 cores, full clocks, ~15 W\n#   MAXN        — 6 cores, moderate clocks, ~10 W\n#   8W_6CORE    — 6 cores, capped at 8 W\n#   8W_4CORE    — 4 cores, capped at 8 W\n#   5W_2CORE    — 2 cores, capped at 5 W (good for overnight idle)\n\nsudo nvpmodel -m MODE_ID\n```\n\nFor an agent running batch jobs overnight, `8W_4CORE`\n\nsaves power without seriously hurting response time. Switch to `MAXN_SUPER`\n\nduring interactive sessions.\n\n**Pro tip:** Add `sudo nvpmodel -m 1`\n\nto your cron agent's startup script so the box drops to low power between scheduled runs.\n\nWith 7.4 GB total and ~2 GB eaten by the OS, you have about 5 GB for the agent + browser + tools. This fits comfortably for:\n\n- Hermes Agent runtime (~200 MB RSS)\n- Chromium with 2–3 tabs (~800 MB RSS)\n- Python tool processes (~100–300 MB each)\n- Local inference? Not really — LLMs need more VRAM. Hermes Agent uses a remote API.\n\nEnable swap if you haven't:\n\n```\nsudo fallocate -l 4G /swapfile\nsudo chmod 600 /swapfile\nsudo mkswap /swapfile\nsudo swapon /swapfile\necho '/swapfile none swap sw 0 0' | sudo tee -a /etc/fstab\n```\n\nThe NVMe-backed swap makes it usable (not SD-card slow), but avoid depending on it for hot paths.\n\nHermes Agent has a built-in cron system. Here's the pattern we use for scheduled technical content:\n\n```\n# Example: ~/.hermes/config.yaml snippet\ncron:\n  jobs:\n    - expression: \"0 6 * * *\"\n      task: \"morning_briefing.md\"\n      delivery:\n        platform: telegram\n      timeout_minutes: 30\n\n    - expression: \"30 */4 * * *\"\n      task: \"web_monitor.md\"\n      delivery:\n        platform: telegram\n      timeout_minutes: 15\n```\n\nThe agent's cron scheduler runs as a background daemon (`hermes cron start`\n\n). Check job status with:\n\n```\nhermes cron list\n```\n\nThe Orin Nano's GPU isn't great for WebGL/Canvas-heavy pages — screenshots work fine, but don't expect smooth 60 FPS page renders. That's fine for:\n\n- ✅ Taking page screenshots for QA/briefing\n- ✅ Filling forms and clicking buttons\n- ✅ Extracting text content and DOM state\n- ✅ Login flows with cookies stored locally\n- ❌ Video-heavy sites at full resolution\n- ❌ CAPTCHA-reliant flows (you'll get flagged anyway on headless)\n\nUse `--disable-gpu`\n\nand `--disable-software-rasterizer`\n\nflags in Playwright to avoid GPU-related crashes:\n\n```\nbrowser = await playwright.chromium.launch(\n    headless=True,\n    args=[\n        \"--disable-gpu\",\n        \"--disable-software-rasterizer\",\n        \"--no-sandbox\",\n    ]\n)\n```\n\n| Path | Use | Recommended |\n|---|---|---|\n`/` |\nOS + snaps | 32 GB |\n`/home` |\nAgent data, models, repos | ≥ 128 GB |\n`~/.hermes/` |\nHermes Agent home (config, logs, sessions) | Auto |\n`~/.hermes/logs/` |\nAgent logs (auto-rotate) | Monitor size |\n`~/.hermes/sandboxes/` |\nEphemeral sandbox directories | On tmpfs if RAM permits |\n\nMove the sandbox to tmpfs if you have spare RAM:\n\n```\n# /etc/fstab entry\ntmpfs /home/clawbox/.hermes/sandboxes tmpfs defaults,noatime,nosuid,size=512M 0 0\n```\n\nThis speeds up ephemeral file ops and reduces NVMe wear.\n\nThe built-in Realtek WiFi (RTL8822CE) works but is mediocre at range. For a 24/7 agent:\n\n**Use Ethernet** if at all possible (avoid driver quirks and reconnects)**WiFi 5 GHz** if Ethernet isn't available (the 2.4 GHz band has more interference)- Add a\n`wlan0`\n\nreconnect cron if using WiFi:\n\n```\n# /etc/cron.d/wifi-watchdog\n* * * * * root ping -c1 -W2 8.8.8.8 >/dev/null || iwconfig wlan0 power off\n```\n\n**Hermes Agent runs fine on Orin Nano** — the constraints are memory and thermal, not CPU. The 6 Cortex-A78AE cores handle model API calls, tool execution, and file operations with room to spare. The GPU is a non-factor for this use case unless you're doing on-device inference (which you probably shouldn't on 8 GB anyway).\n\n**The biggest bottleneck** is the *agent's API round-trip*, not the hardware. Running a local LLM on this box is impractical — the unified memory is too small. But as a remote-API agent runner, it's excellent.\n\n**Daily operation** is dead simple: the box sits on a shelf, boots Ubuntu, launches Hermes Agent's cron daemon, and executes scheduled tasks. We SSH in maybe once a week for maintenance.\n\n[ClawBox](https://clawbox.tech/)— A turnkey self-hosted AI agent appliance based on this exact Jetson Orin Nano + Hermes Agent stack, in a compact enclosure ready to ship to your desk.[Hermes Agent](https://hermes-agent.nousresearch.com/)— The open-source AI agent runtime with MCP, browser tools, kanban, cron, and gateway integrations.", "url": "https://wpnews.pro/news/hermes-agent-on-nvidia-jetson-orin-nano-practical-setup-guide-config-and-health", "canonical_source": "https://gist.github.com/KrasimirKralev/1dd46951659a46666b20ff22e52f193a", "published_at": "2026-06-12 07:35:07+00:00", "updated_at": "2026-06-25 19:12:38.024312+00:00", "lang": "en", "topics": ["ai-agents", "developer-tools", "ai-infrastructure", "ai-products"], "entities": ["NVIDIA", "Jetson Orin Nano", "Hermes Agent", "JetPack", "Ubuntu", "Chromium", "Playwright", "NVMe"], "alternates": {"html": "https://wpnews.pro/news/hermes-agent-on-nvidia-jetson-orin-nano-practical-setup-guide-config-and-health", "markdown": "https://wpnews.pro/news/hermes-agent-on-nvidia-jetson-orin-nano-practical-setup-guide-config-and-health.md", "text": "https://wpnews.pro/news/hermes-agent-on-nvidia-jetson-orin-nano-practical-setup-guide-config-and-health.txt", "jsonld": "https://wpnews.pro/news/hermes-agent-on-nvidia-jetson-orin-nano-practical-setup-guide-config-and-health.jsonld"}}