Test driving Hermes what am I missing?
I have done my fair share of AI work, custom setups with vercel sdk, N8N, windmill, custom nestjs
Have Claude openai perplexity accounts and subs to open router and deepseek. Done a lot of skill and MCP creations
I was missing a setup that runs autonomous and does hard grunt work for me.
Test driving Hermes but a bit disappointed, I am a fan of opensource and appropriate this project but I am still missing all the hype.
- the Kanban setup I like the idea, but can't delete a task? No dropdown to select a profile?
- weird setup with profiles which don't really act as subagents?
- the UI sucks, really hard to read fonts and colors?
- no way to expand the logs per Kanban task properly? I am scrolling in a tiny window
- I like the gateway thing, but using telegram for a while.. it doesn't fit the purpose, shouldn't this project have a dedicated (web) app for managing the setup?
Within 10 minutes I had quirks/bugs and patches which don't survive updates
Watched some YT videos but it doesn't really seem to click for me in the ergonomics department. I feel like I am missing something obvious. I know exactly what I want but it seems a different thinking paradigm.
Again not hating, hope to get some tips or closure haha.