Same Prompt, Very Different Automations

Same Prompt, Very Different Automations

🔗 Link do vídeo: https://www.youtube.com/watch?v=VruF9Z6WrGg
🆔 ID do vídeo: VruF9Z6WrGg

📅 Publicado em: 2026-07-03T15:20:02Z
📺 Canal: n8n

⏱️ Duração (ISO): PT6M35S
⏱️ Duração formatada: 00:06:35

📊 Estatísticas:
– Views: 697
– Likes: 40
– Comentários: 7

🏷️ Tags:

We gave Claude Opus 4.8 the same prompt in two environments: the Hermes personal AI assistant and n8n. The task: an X (Twitter) profile tracker that sends Telegram alerts when tracked accounts post something relevant to your niche.

Same agent, same prompt. In Hermes, the output is generated code you manage through further prompting. In n8n, it's three documented workflows you can open, inspect and edit. This video compares Hermes Agent, OpenClaw and n8n, shows both builds end to end, and breaks down what each approach costs.

🎬 IN THIS VIDEO

🔹 How Hermes Agent, OpenClaw and n8n compare: agent interface vs automation UI
🔹 Building the bot in Hermes from a single prompt, then testing it in Telegram
🔹 What "vibe coding" an automation abstracts away: architecture, database security, and API costs you stop seeing
🔹 The same build in n8n using Claude Code, the n8n MCP and the official n8n skills
🔹 Cost breakdown: infrastructure, tokens, the X API, and the long-run cost of needing an LLM for every change
🔹 Using them together: managing n8n workflows from Telegram, Discord, Slack or WhatsApp

⏱️ TIMESTAMPS

00:00 Intro & what each platform is
01:08 Building the bot in Hermes
02:14 The trade-off: maintainability & observability
03:04 Building the same bot in n8n
05:16 Costs & how these tools work together

🔗 TOOLS & RESOURCES

n8n → https://n8n.io

Official n8n skills for Claude Code → https://github.com/n8n-io/skills

n8n MCP server → https://github.com/czlonkowski/n8n-mcp

Liam's video on MCP usage → https://www.youtube.com/watch?v=FXtjGGAazT0

💬 Are personal AI assistants competitors to n8n, or companions? Tell us in the comments.