Hermes Agent, AI
Hi fam. Anyone using Hermes for Fluentcommunity (automation, management, marketing, content, workflows)?
Looking for alternative solution's besides the expensive and unpredictable subscription plans/fees for llm, tools, API, MCP connectors, skills (Claude, OpenClaw, etc)..
#HermesAgent is an open-source, self-hosted AI agent framework developed by Nous Research and released in February 2026, autonomous agents with persistent memory and cross-platform messaging.
I've been using different agentic frameworks since around January (primarily OpenClaw), and overall the experience has been excellent.
That said, if you're not using top-tier models, things can get messy very quickly. Poor model quality can easily break workflows, create unpredictable behavior, and introduce serious security, legal, and operational risks, especially if you don't have proper protocols and safeguards in place.
The Fluent stack is honestly fantastic for this new agentic era (I am sure Shahjahan Jewel may be cooking something related to this).
It’s relatively lean, flexible, and powerful. But you still need to be extremely careful with model quality, prompting, permissions, and system design.
And above all: before implementing any AI or agentic workflow in production, I strongly recommend conducting a proper privacy, security, and data protection assessment.
GDPR fines are no joke. And starting in August, the regulatory landscape in the EU will become even stricter in this area, particularly if you're processing EU citizens’ data or exposing AI/agent-generated content to EU audiences.
My conclusion after months of testing:
do not cut corners on model quality.
The technical, privacy, and security risks simply aren’t worth the savings.
I’ve also spent quite a bit of time experimenting with local models and self-hosted setups. They’re improving fast, but in my experience, the quality gap is still significant for serious production use cases.
Hope it helps!
Jorge de los Reyes I sometimes wondered if you were an AI 😜
You seem to be a wealth of knowledge!
Im starting to look into this at the moment too. Ross FCA has got one working on his community and I read rumours of a plugin in the works.
Ive been toying with OpenWebUI, Ollama & N8n for this but am yet to fully sink me teeth into it yet.
Jorge de los Reyes could you possibly be able to give a brief outline on how you set it up?
Luke HubCo 🙈 That combo looks really good.
If you want to go deeper, I’d probably look at Ollama and n8n as building blocks for agent workflows.
One thing I’ve noticed is that, once agents start handling more of the actual operations, traditional SaaS dashboards become less central. That said, I still like having a Mission Control-style dashboard for visibility: what is running, what failed, what needs human review, and where the risks are.
My main advice would be: security first, convenience second.
Before setting up anything serious, I’d go deep into the docs of the framework you choose, start with everything closed by default, and only open features as you really need them. Also document what each agent can access, what it can do, and where human approval is required. That matters both operationally and legally.
The exciting and slightly scary part is that this is now outrageously easy to set up technically. But the easier it gets, the more important it is to be conservative with permissions.
If you get stuck on the technical setup, happy to take a look!
Just got a mac-mini today to start playing with Hermes agent. Would be interested in doing some testing.
Richard Roy hows that working out? #Hermes #MacMini
thanks for the comments, lets keep each other updated along the AI journey..
