I will build jarvis ai assistant hermes agent dashboard auto post n8n mcp claude code


About this gig
build jarvis ai assistant hermes agent dashboard auto post n8n mcp claude code
You approve every post by hand and you still cannot see what your agents did while you were away
- Build your jarvis ai assistant on hermes agent or openclaw with claude code as the brain
- Set up a dashboard showing task status, token spend, and what each agent did overnight
- Auto post to x, linkedin, instagram, facebook, and youtube from one n8n workflow
- Schedule your content with cron so drafts queue, publish, and report back on their own
- Add human approval in telegram or whatsapp before anything goes out to your audience
- Draft your captions and hooks with claude, then repurpose one piece across every channel
- Write custom mcp servers so your assistant reaches your cms, drive, or internal tools
- Store your posted history and performance in obsidian or notion for later reference
- Add memory md, user md, mem0, or zep so your assistant keeps your tone across posts
- Deploy on hetzner, contabo, or hostinger vps with docker and tailscale access
No setup, partial setup, or something broken, I can work with any of them
Message me your channels and what you want posted
Get to know Kathryn B
I fix and build Hermes Agent trading bots and AI automation systems
- FromGermany
- Member sinceJul 2026
- Avg. response time1 hour
Languages
English, Spanish, German, Italian
FAQ
Can it post to more than one platform from a single draft?
Yes. I build one n8n workflow that reformats a single draft for x, linkedin, instagram, facebook, and youtube, adjusting length and hashtags per platform instead of you rewriting the same idea five separate times.
Does anything publish without me seeing it first?
Yes or no, your choice. I add an approval step in telegram or whatsapp where you tap to approve, edit, or reject each queued post, or you can let low risk content publish automatically on schedule.
What does the dashboard actually show me?
Yes, real detail. The dashboard shows queued and published posts, task status per agent, token spend by model, and errors that need attention, so you can see what hermes agent and claude code did while you were away.
Will the posts sound like me or like generic ai output?
Yes, like you. I store your tone, phrasing, and past posts in obsidian or notion and load them through memory md, user md, mem0, or zep, so claude drafts in your voice instead of default assistant phrasing.
Can it pull content from my own notes instead of inventing topics?
Yes. I connect your obsidian vault or notion database through mcp, so the assistant drafts from your existing notes, articles, and ideas rather than generating filler topics you never planned to publish.
What if a platform api changes and posting breaks?
Yes, this happens. I build retry handling and failure alerts into the n8n workflow, so you get notified in telegram when a post fails instead of finding out days later that nothing published
Where does the whole thing run?
Yes, on your own server. I deploy to hetzner, contabo, hostinger, digitalocean, or railway with docker, and add tailscale so you reach the dashboard remotely without exposing it to the open internet.

