Alex Finn’s Hermes Agent Advice
A dense, timestamp-backed briefing document covering Alex Finn videos from the past 3 months that are directly about Hermes Agent or meaningfully transferable to Hermes-style personal agent workflows.
Executive summary
- Alex’s strongest Hermes advice is not a single feature; it is a system: memory first, visible queues second, recurring/background work third, and role-specific agents only when the role truly needs separate memory/tools/models.
- The repeated setup pattern is: brain-dump context, ask the agent to reverse-prompt workflows from that context, schedule a small number of useful recurring tasks, then audit whether memory/skills/cron actually updated.
- Hermes Desktop matters mostly as a control surface: sessions, artifacts, skills/toolsets, cron, profiles, and API keys are easier to inspect than in a chat-only workflow.
- Kanban is the cleanest operating layer for agent work: human tasks stay with the human; agent-suitable tasks move to triage, get broken down/enriched from memory, then run through profiles/subagents.
- For cost and reliability, Alex repeatedly points toward a brain-and-muscles architecture: strong cloud model for orchestration/review, cheaper Codex/local/specialized workers for execution.
- The best reportable workflows are practical and recurring: morning priority checks, nightly micro-app/task crons, daily memory logs, research briefs, system monitors, and periodic code-quality reviews.
Ranked takeaways
#1🧠 Build memory first
Brain-dump identity/goals/workflows; keep daily logs and memory wiki pages; verify memory is actually written.
#2🔁 Reverse-prompt your system
Ask Hermes what workflows, cron jobs, dashboards, and subagents it should create based on your context.
#3🗂️ Use Kanban as the agent handoff layer
Put AI-doable work in triage, let Hermes enrich/break down tasks, then assign to profiles/subagents.
#4⚡ Run background and cron work deliberately
Use /background for parallel tasks; use Cron UI to verify recurring jobs; make outputs visible.
#5👥 Keep profiles/subagents role-specific
Use profiles when memory/tools/models/personality differ; start with one or a few stable roles.
#6💸 Route by cost and capability
Use a strong orchestrator for judgment; cheaper/local/Codex-style workers for execution, research, scraping, and coding.
#7🛡️ Keep secrets and risky actions controlled
Use API-key settings, not chat logs; route risky approvals/spending back to the user.
#8🖥️ Make agent work visible
Use Desktop, Artifacts, Mission Control, task boards, timelines, and notifications so work is not hidden in chat.
#9📚 Treat skills as reusable procedures
Let Hermes capture repeated workflows as skills, but audit and prune unused skills/toolsets.
#10✅ Verify outputs
Have agents test apps with browser/computer use, recap long sessions, and produce evidence links/timestamps.
Video-by-video notes
Newest first. Each timestamp links to the source video at that point. “Medium” and “Low” videos are included for transferable workflow context or brief Hermes-specific mentions.
Video index
- 2026-06-09 — Claude Fable 5 just dropped and I'm speechless...
- 2026-06-03 — Hermes Agent just WON (Hermes desktop app)
- 2026-05-28 — Claude Opus 4.8 actually blew my mind...
- 2026-05-26 — Hermes Agent is the greatest AI tool ever made. Here's how to set it up
- 2026-05-22 — 6 Hermes Agent use cases I promise will change your life
- 2026-05-19 — Hermes just got 10x better...
- 2026-05-17 — Claude Code is 1000x better when you use this tool
- 2026-05-13 — Hermes Agent powered by local models on the DGX Spark is basically magic
- 2026-05-09 — Hermes Agent is blowing me away...
- 2026-05-05 — Hermes Agent might have just killed OpenClaw
- 2026-04-30 — ChatGPT 5.5 Codex is the greatest AI coding tool ever. Here's how to use it
- 2026-04-28 — Hermes Agent w/ ChatGPT 5.5 is literally magic
- 2026-04-25 — ChatGPT 5.5 Codex: I can't believe they did this...
- 2026-04-22 — The best Claude Design workflow you’ll ever see…
- 2026-04-21 — OpenClaw Full Tutorial: Set up your first AI employee!
- 2026-04-18 — The creator of Claude Code just revealed 7 secrets to using Claude Code (Opus 4.7)
- 2026-04-15 — Claude Code for Desktop is the BEST way to build apps with AI EVER (full tutorial)
- 2026-04-13 — You NEED to set up a multi agent team with OpenClaw and Hermes
- 2026-04-09 — OpenClaw + Obsidian gives you super powers
- 2026-04-05 — Anthropic just blocked OpenClaw. Here’s what you need to do immediately
- 2026-03-31 — MASSIVE CLAUDE CODE LEAK! THIS CHANGES EVERYTHING!
- 2026-03-31 — Did Hermes Agent just kill OpenClaw? (full guide)
- 2026-03-24 — Why you NEED to be running local AI models (FULL beginners guide)
- 2026-03-22 — 5 OpenClaw use cases that will actually improve your life...
- 2026-03-17 — The only OpenClaw tutorial you’ll ever need (March 2026 edition)
- 2026-03-12 — How to build an army of OpenClaw agents
1. Claude Fable 5 just dropped and I'm speechless...
2026-06-09LowLow — mostly Claude/Fable, but includes agent workflow patterns and a brief Hermes/OpenClaw teaser.
- 01:45Use loops/autonomy to move from checking whether an agent did work to checking whether it is doing the right work.
- 02:08Treat stronger models as thought partners: ask how they would approach or change the plan before building.
- 05:40Use an advanced planning/interview mode before building complex productivity systems.
- 07:45Paste the generated spec into a long-running goal-style workflow so the agent loops until success criteria are met.
- 11:19Adjacent workflow: loop over Linear tasks periodically so agents can pick up new work automatically.
2. Hermes Agent just WON (Hermes desktop app)
2026-06-03HighHigh — direct Hermes Desktop walkthrough.
- 00:54Use Desktop sessions/threads for each area of life and pin important ones so scheduled sessions do not bury them.
- 03:07Use Artifacts as a second brain for links, files, images, and generated outputs.
- 04:43Disable unused skills/toolsets to reduce context, tokens, noise, and cost.
- 06:13Use the Cron UI to verify, manually create, test, and pause scheduled jobs.
- 07:27Use profiles only when roles need different memories, skills, personalities, tools, or models; otherwise use subagents.
- 10:22Store API keys in the settings/API-key UI instead of pasting secrets into chat logs.
3. Claude Opus 4.8 actually blew my mind...
2026-05-28LowLow — mostly Claude/Claude Code; includes model-switching caution for Hermes/OpenClaw.
- 05:29Switch Claude Code tasks to a new model first, but do not immediately force Hermes/OpenClaw onto it.
- 06:23Wait for official Hermes/OpenClaw support before changing the agent model, because brand-new model IDs can crash agent setups.
- 07:37Stay focused while agents work; do not prompt an agent and then waste the recovered time.
- 09:28Use remote-control/mobile monitoring for coding agents when away from the desktop.
HighHigh — full Hermes setup/tutorial.
- 01:51Think of Hermes as a 24/7 AI employee that learns goals, workflows, and preferences.
- 08:33Install from the Hermes website command, then choose model and messaging channel; Alex favors Telegram.
- 09:28Model tradeoff: Claude/Anthropic best but expensive; ChatGPT subscription is a moderate option; local/open models can be cheaper but less capable.
- 13:44First task: brain-dump who you are, what you are working on, and goals so Hermes can save useful memories.
- 15:23Schedule a nightly 2am proactive cron to build a micro-app/UI/system that advances your goals.
- 21:02Use Kanban triage: add goals/tasks, let Hermes split subtasks, move them to to-do, and assign subagents.
- 29:41Memories and skills are local markdown files; Hermes can review execution and improve future repetitions.
5. 6 Hermes Agent use cases I promise will change your life
2026-05-22HighHigh — direct Hermes use cases.
- 00:49Use /goal for long-running tasks, but meta-prompt first so the goal is detailed rather than vague.
- 03:23Morning Kanban workflow: put AI-doable tasks into triage while you do human-only work.
- 06:29Use browser/computer control for competitor app research: inspect site, console, features, and stack, then produce a report.
- 08:58Build a memory wiki with subjects and daily logs so you and Hermes can review past work.
- 10:36Use Tailscale to let Hermes administer/test across trusted devices.
- 12:51Schedule a morning priority prompt: ask the #1 priority, propose support tasks, and update memories.
6. Hermes just got 10x better...
2026-05-19HighHigh — direct Hermes feature update.
- 00:37Improved session recall lets Hermes retrieve prior sessions by date/time/topic without bloating current context.
- 02:21/background lets Hermes run tasks in parallel while remaining available for normal chat.
- 05:22Native Codex CLI support lets Hermes spin up Codex as a coding worker.
- 07:02Computer use lets Hermes inspect and control apps remotely, e.g. calendar or Notion workflows.
- 10:43Auto-Kanban task generation breaks triage goals into subtasks assigned to agents/subagents.
7. Claude Code is 1000x better when you use this tool
2026-05-17MediumMedium — coding-agent board workflow transferable to Hermes/Kanban.
- 00:35Use a task board as an external second brain for coding agents.
- 05:36Before coding, have the agent create detailed issues/projects for the app.
- 09:02Acceptance criteria, priorities, and dates reduce drift in long coding runs.
- 14:56Multiple coding agents can coordinate through the same board rather than manual context syncing.
- 16:46Put durable process rules in AGENTS.md/CLAUDE.md: read issue specs, meet acceptance criteria, avoid unrelated refactors.
HighHigh — direct Hermes local-model setup.
- 01:27Local models make 24/7 Hermes work more private and effectively free after hardware/electricity.
- 05:40Use cloud-powered Hermes on the main computer to configure the headless/local model box.
- 06:58Install Tailscale so Hermes can control the local machine over a private network.
- 09:55Prompt Hermes to install a local model such as Qwen and validate it with a simple chat frontend.
- 13:37Create a separate Hermes profile connected to the local model.
- 16:25Use local Hermes for recurring research, transcript repurposing, self-improvement from videos, and lightweight coding.
9. Hermes Agent is blowing me away...
2026-05-09HighHigh — direct Hermes setup/use cases.
- 02:42Hermes self-improves by creating/updating skills when it encounters new or repeated tasks.
- 04:31Hermes supports swarms/profiles; ask it to build a new profile for specialized agents.
- 05:05Setup flow: terminal command, Telegram channel, and model choice; Opus best if budget allows.
- 10:03Brain-dump career, goals, interests, and workflows into Hermes after setup.
- 11:28Reverse-prompt Hermes for workflows based on what it knows about you.
- 13:08Starter cron: daily AI news plus new AI tool recommendations based on your workflows/memory.
10. Hermes Agent might have just killed OpenClaw
2026-05-05HighHigh — direct Hermes dashboard/Kanban guide.
- 03:34Open the Hermes dashboard with `hermes dashboard`; use Kanban for many parallel tasks.
- 04:44Kanban flow: triage → memory enrichment → ready → assigned execution.
- 08:08Create a cron that checks Kanban every 10 minutes, expands triage tasks, and executes ready tasks.
- 08:54Use a second admin/librarian profile for recurring Kanban management so the main agent is not blocked.
- 13:35Profiles can have separate memories, skills, env vars, and roles; baseline: main, coding, research, admin.
- 16:21Lower compression threshold to about 0.5 and use curator reports to reduce memory/skill bloat.
MediumMedium — coding-agent workflow transferable to Hermes.
- 02:11Generate UI options before coding so design direction is clearer.
- 06:31Run multiple sessions/agents in parallel for app, marketing, research, or other tracks.
- 07:16Install specialized skills/plugins such as browser use, computer use, and video/design capabilities.
- 12:17Use browser/annotation feedback and have the agent test itself.
- 19:03Schedule recurring code-quality checks over recent commits.
12. Hermes Agent w/ ChatGPT 5.5 is literally magic
2026-04-28HighHigh — direct Hermes + ChatGPT setup/use cases.
- 01:01Hermes can act as a 24/7 computer employee: build apps, send email, check calendars, move files, and act on the computer.
- 01:23Self-improvement through skills is a core advantage; repeated usage makes future tasks better.
- 03:53ChatGPT 5.5 via OAuth is positioned as a lower-cost/high-limit Hermes brain versus Claude Opus.
- 04:57Use Telegram BotFather setup to communicate with Hermes from multiple devices.
- 06:35Use mobile Telegram prototyping and /steer to correct an active task mid-run.
- 10:42Use cron for daily research or checking another agent every few hours.
13. ChatGPT 5.5 Codex: I can't believe they did this...
2026-04-25MediumMedium — coding-agent practices transferable to Hermes.
- 04:20Organize by project and run backend/frontend/planning sessions in parallel.
- 05:33Enable computer use so the agent can test apps and interact with the browser/computer.
- 08:10Start with image-generated interface options before implementation.
- 13:43Ask the agent to click through and test all app features while you do other work.
- 16:19Connect a project-management layer so multiple agents share issues and details.
14. The best Claude Design workflow you’ll ever see…
2026-04-22MediumMedium — design workflow transferable to Hermes coding/design tasks.
- 02:43Separate cheap visual exploration from high-fidelity design/prototyping.
- 03:00Start with inspiration images/logos/styles and ask for several design directions.
- 13:05Generate a brand guide/prompt after choosing the design direction.
- 15:14Build a reusable design system before coding full interfaces.
- 20:00Hand the design system/prototype to the coding agent for implementation.
15. OpenClaw Full Tutorial: Set up your first AI employee!
2026-04-21HighHigh — OpenClaw tutorial with direct Hermes comparison and transferable setup workflows.
- 13:53Use messaging surfaces intentionally: Telegram for daily chat; group topics for separate contexts.
- 19:49Brain-dump identity, business, goals, and workflows to personalize the agent.
- 20:36List manual computer tasks for a day, then ask the agent which workflows it can automate.
- 23:00Use a project board/Linear-style workflow to assign tasks, branch, and PR code.
- 30:50Hermes is described as faster/more token-efficient, while OpenClaw memory was stronger in his experience; use both if helpful.
- 32:42Two-agent reliability: if one agent breaks, the other can inspect configs/code and fix it.
MediumMedium — coding-agent operating practices transferable to Hermes background/subagent work.
- 01:15Use higher-autonomy modes to reduce repeated permission prompts when safe.
- 02:40Front-load full task information at the start of a session.
- 04:43Use Hermes/OpenClaw-style agents for on-the-go prototypes and tooling; use coding agents for deep production coding.
- 05:56Ask for recap after long-running sessions so you can review unattended work.
- 08:04Turn on notifications/hooks so long work pings you when complete.
MediumMedium — desktop coding-agent workflow transferable to Hermes Desktop/Kanban supervision.
- 00:19Project-oriented desktop UI helps supervise many sessions with attention indicators.
- 02:26Keep multiple local coding sessions on separate app areas to avoid conflicts.
- 06:57Use accept-edits mode for speed, plan mode for reasoning/approval.
- 10:11Use routines/scheduled tasks for nightly code review and bug fixing.
- 15:02Tool split: coding agent for production app coding; Hermes/OpenClaw for prototypes, mission-control apps, and agent tooling.
HighHigh — direct OpenClaw + Hermes multi-agent setup.
- 00:31Use two agents for reliability: when one breaks, the other can diagnose and fix it.
- 02:28Run Hermes as a lighter/faster assistant or monitor while another agent handles heavier main work.
- 03:53Allocate models by role: strong expensive model for main/orchestrator, cheaper model for monitoring/execution.
- 07:57Planner/builder/reviewer workflow: strong agent plans, Hermes executes, strong agent reviews.
- 12:11Schedule Hermes cron jobs to monitor systems built by another agent and alert on failures.
- 13:44Use shared memory folders so agents have private logs plus shared decisions, mistakes, plans, and context.
19. OpenClaw + Obsidian gives you super powers
2026-04-09HighHigh — memory/wiki workflow explicitly useful for Hermes too.
- 00:00An Obsidian-backed memory system can improve both OpenClaw and Hermes memory.
- 00:26Use daily logs that capture discussions, tasks, and work done.
- 00:58Keep a mistakes file so corrections become future behavior improvements.
- 01:18Use an agent-shared workspace for multi-agent context handoff.
- 03:56Keep large memory in files checked on demand, not injected into every prompt.
- 09:04Verify the agent is writing memories; if not, add the memory rules to AGENTS.md/agent instructions.
MediumMedium — model-routing/cost strategy transferable to Hermes.
- 01:24Use a brain-and-muscle architecture: strong orchestrator plus cheaper executors.
- 03:53For average hardware, keep a strong cloud orchestrator and route coding/other tasks to cheaper cloud/subscription models.
- 07:53For strong hardware, use local models for coding, scraping, and data collection.
- 13:56The orchestrator chooses the right cheaper muscles to preserve output quality while reducing cost.
- 14:36Evaluate agent spend as ROI on 24/7 digital labor, not as a normal entertainment subscription.
21. MASSIVE CLAUDE CODE LEAK! THIS CHANGES EVERYTHING!
2026-03-31MediumMedium — future coding-agent patterns transferable to Hermes expectations.
- 00:55Proactive mode idea: agent proposes and builds follow-on tasks rather than only executing commands.
- 03:10Dream/overnight mode: agent thinks about recent work and what to build next while you sleep.
- 04:11Smart auto approvals sit between YOLO and constant manual prompts.
- 05:26Risky approvals can be routed to messaging apps for user confirmation.
- 05:53Agentic payments could remove blockers like buying hosting/accounts/templates, but would need explicit spending controls.
22. Did Hermes Agent just kill OpenClaw? (full guide)
2026-03-31HighHigh — direct Hermes guide/comparison.
- 01:52Hermes is described as lightweight and performant compared with OpenClaw.
- 02:21Hermes can perform a complex briefing workflow and capture it as a reusable skill.
- 03:21Hermes is transparent about tool calls, websites visited, snapshots, and what it is doing.
- 03:42Hermes creates a Hacker News daily AI briefing skill, updates a cron job, and delivers via Telegram.
- 07:22Obsidian-style memory improved both Hermes and OpenClaw memory.
- 08:45Use Hermes and OpenClaw side by side; one can spin up or fix the other.
- 12:38Model choices: Claude most agentic, ChatGPT cheaper/strong, local models possible for advanced users.
MediumMedium — local-model strategy transferable to Hermes local/open-model profiles.
- 01:07Cloud models are expensive, less private, and hard to scale for 24/7 agents.
- 03:19Local models are private, customizable, offline-capable, and effectively free after hardware.
- 04:52Local models unlock 24/7 scraping, writing, coding, and recurring work without token bills.
- 11:31Choose model size based on hardware; mentions Qwen, Nemotron, and MiniMax as options.
- 13:19Ask the agent to inspect hardware and Hugging Face before recommending local models.
- 15:30Use cloud orchestrator + local worker muscles for coding, research, scraping, and writing.
24. 5 OpenClaw use cases that will actually improve your life...
2026-03-22MediumMedium — OpenClaw use cases map to Hermes cron/memory/workflow patterns.
- 00:23Daily memory tracker records discussions, coding, shipped work, and projects.
- 02:18Prompt pattern: summarize every discussion into daily logs and refer back for project/day questions.
- 03:22Trending content alerts can run through Discord channels for rapid response.
- 05:54Use agents to vibe-code internal microapps/tools; use dedicated coding tools for serious consumer apps.
- 07:11Reverse-prompt the agent for app ideas based on your workflows and memory.
- 08:28R&D lab workflow: multiple models review recent work, debate ideas, and produce a memo.
- 12:31Overnight employee cron: review goals and do one task that moves them forward.
MediumMedium — comprehensive agent-workflow tutorial transferable to Hermes.
- 02:37Agent framing: give goals and get results; agent iterates and self-improves.
- 14:18Use Telegram for daily chat and Discord for structured multi-channel workflows.
- 19:23Schedule recurring research briefs via cron.
- 24:06Brain dump goals/preferences/workflows, then ask for custom automations.
- 25:44Use Discord as second brain with channels for briefs, drafts, alerts, and digests.
- 28:12Do not blindly install third-party skills; have the agent inspect and recreate safe versions.
- 32:47Build Mission Control for calendar, memory, docs, office/team views, and custom tools.
26. How to build an army of OpenClaw agents
2026-03-12MediumMedium — subagent/org design transferable to Hermes multi-agent operations.
- 01:22Start with one subagent rather than many; scale only after a clear use case.
- 02:42Choose the first subagent based on daily work: coding, writing, or research.
- 03:21Route tasks to the right model: strong orchestrator, coding model, cheap/local research/writing models.
- 05:54Let the orchestrator configure subagents instead of manually editing configs.
- 07:02Subagents enable parallel work, lower cost, and separate context windows.
- 08:21Maintain an org/team screen listing each agent, model, and responsibility.
- 15:21Reverse-prompt for org design: ask which subagents should exist based on your goals/workflows.
Appendix: method and files
Source channel: Alex Finn YouTube videos page. I refreshed metadata for the recent channel videos, limited the research window to the past 3 months, reused cached transcripts when present, fetched missing transcripts, and extracted timestamped points from transcript text.
Local working files: videos.jsonl, relevant-videos.jsonl, video-key-points.jsonl, extracted-tips.jsonl, and transcripts/<video_id>.txt.
No included transcript was unavailable. Transferable videos are not represented as direct Hermes product claims.