Hermes Agents · Buzz Communities · Local AI · Johannesburg

Your business,
running on Hermes, Buzz
and your own local AI.

Three ways to put AI to work. Hermes agents run your admin while you sleep — answers at 2am, documents chased, work shipped. Buzz communities give your team and clients one room with an agent in it. Local AI engines put a model on your own hardware — sized to your machine, no per-token credit bills. From R2,500 once-off.

Start with a discovery call See how it works

R2,500.

Hermes agent — once-off

R2,500.

Buzz community — once-off

R7,500.

Local AI engine — from

1 day.

Typical install & handover

24/7.

They work when you don't

Engraving of the winged sandal of Hermes — Hermes

Engraving of a beehive with bees — the Buzz community — Buzz

Setting up agents & communities for Financial advisory Property & real estate Professional services Retail & wholesale Mining & logistics Hospitality

Hermes Agent Setup

We install, train, and support production-ready Hermes agents on your machine — skills, memory, MCP, and real work output.

Full install & configuration
Baseline skills & memory on
MCP connections from R1,000
Recorded training & handover

R2,500 once-off →

Buzz Community Setup

We create your community on Buzz, brand it, set channels, and wire a Hermes-powered agent that answers @mentions and DMs.

Community created & branded
Channels & roles configured
Hermes-powered agent in the room
@mention replies & DM routing

R2,500 once-off →

Local AI Engine

We set up a model on your hardware — sized to your machine — so your AI runs without per-token credit bills.

Hardware audit & sizing
Model installed & tuned
Local API + desktop UI
Hermes integration add-on

R7,500 from →

01 · Hermes agent setup

A production Hermes agent
on your machine — day one.

We install, configure, and hand over a working Hermes agent: models, baseline skills, memory, channels — plus a 90-minute training session so your team owns it.

See what's included

8GENTS · HERMES SETUP

# basic local agent setup — R2,500 once-off

hermes setup

Hermes runtime installed

hermes skills install ./your-business

6 baseline skills ready

hermes channels connect whatsapp email

Channels live

hermes memory verify

Persistent memory on — local only

# handover session: 90 min, your team, recorded

Agent ready. It works while you sleep.

02 · Buzz community setup

A Buzz community for your business — with a Hermes agent in the room.

Buzz is a Nostr-based workspace where humans and AI agents share one room. We set up your community, channels, and a Hermes-powered agent that answers @mentions and DMs.

8GENTS · BUZZ SETUP

# Buzz community creation — R2,500 once-off

buzz community create your-business

Community live at your-business.communities.buzz.xyz

buzz channels add general announcements clients

3 channels ready

hermes gateway start --platform buzz

Agent wired — answering @mentions & DMs

# handover: community + agent training

Community ready. Room with an agent — yours.

/ 01

Your community, branded

Public or private community on your own relay. Channels set, roles configured, invite-ready.

/ 02

Agent that answers

Hermes-powered agent responds to @mentions and DMs — same memory, same context across channels.

/ 03

Room with a brain

Your team, clients, and AI agent in one space. Automated updates, follow-ups, overnight coverage.

03 · Local AI engine

Your AI, on your hardware.
No per-token bills.

We audit your machine, pick the best model it can actually run, and install it — Ollama or llama.cpp, quantized to your RAM and GPU. Your data stays on your hardware and the only running cost is electricity.

See local pricing

8GENTS · LOCAL AI SETUP

# hardware audit — R7,500 from

8gents audit hardware

16GB RAM · 4GB VRAM — 7B Q4 recommended

ollama pull qwen2.5:7b

Model running — ~5GB RAM, fully local

hermes model set local

Hermes now answers from your hardware

# no per-token bills. ever.

Credits: 0. Running cost: electricity.

04 · How they work together

Hermes + Buzz — one agent, every surface.

Your Hermes agent handles email, WhatsApp, CRMs, and filesystem. Your Buzz community gives your team a shared room with the same agent inside it.

◆

One memory

Client context follows the agent everywhere — email, WhatsApp, Buzz DMs, and @mentions draw from the same memory.

⇄

Every channel

Hermes on WhatsApp + email + web. Buzz for your team. Same agent, same brain, no gaps.

◦

Local & private

Both run on your hardware by default. Your files, memory, and community data stay under your control.

↻

Improves with use

Skills, workflows, and Buzz agent behaviour sharpen as your team uses them — setup is the start.

⌚

Proactive, 24/7

Scheduled scans, overnight reports, and Buzz community updates — your agent checks in with work done.

✋

You stay in control

Approvals on high-value actions. Clear logs of what ran and why, on both Hermes and Buzz.

⚙

MCP tools

Filesystem, CRM, calendars, APIs — Hermes actions reach into Buzz channels and DMs.

◎

Always on

Optional hosting for 24/7 uptime — from R500/month — keeps both Hermes and Buzz alive.

05 · How it works

Discovery call to working agent
or community — in days.

Discovery call

30–45 minutes. We map your highest-pain work and pick the fastest win — Hermes, Buzz, or both.

Free · this week

Setup

We install Hermes on your machine and/or create your Buzz community — models, memory, channels, branding.

Same day

Training & handover

90-minute live session. Your team learns to run the agent and the community — recorded.

Recorded

Scale what works

Add skills (from R500), MCP (from R1,000), custom connectors (from R3,500), Buzz upgrades (from R4,500).

Modular · clear prices

06 · Where it lands

Built for businesses that run on follow-ups.

Keep scrolling

01Financial & advisory

More billable hours, less admin

Meeting prep, ROA drafting, client follow-ups, document assembly — handled between appointments.

02Property & real estate

Every lead answered, every time

Instant WhatsApp replies, viewing scheduling, document chasing, status updates to buyers and sellers.

03Professional services

Onboarding that runs itself

Client intake packs, proposals, invoice follow-up, knowledge not living in one person's head.

04Retail & wholesale

Customers answered at scale

Order queries, supplier nudges, multi-channel support — without extra headcount.

07 · Pricing

Three services. Clear entry. Modular scale.

No retainers to start. No hidden fees. You know the price before we begin.

Hermes agents — setup & skills

DIY vs done for you.

	DIY Hermes	With 8gents
Who owns setup	You + forums	We install & train
Time to useful agent	Often months	Days for basic setup
Business workflows	You invent skills	Scoped skills with you
Stack connections	Trial and error	MCP connection or custom build
Buzz community + agent	Empty community, no replies, trial-and-error	Community created, agent answering @mentions day one
Ongoing help	Community only	Direct support path
Cost shape	Your time + surprises	Clear packages from R2,500
AI running costs	Per-token bills, forever	Owned hardware — no credits, no monthly fees

09 · FAQ

Straight answers.

What exactly do I get for R2,500?

For Hermes: a production-ready agent on your machine — models, skills, memory, channels, 90-minute handover. For Buzz: community created & branded, channels set, Hermes-powered agent answering @mentions and DMs. Same price, different delivery.

Does my data leave my machine?

Core agent and community operations run locally. Your files and memory stay on your hardware. Model API calls follow the providers you choose — we configure that with you.

What's the difference between Hermes and Buzz?

Hermes is your business AI agent — it runs on your machine, connects to email, WhatsApp, CRM, and files. Buzz is your community — a shared room for your team, clients, and the same Hermes agent. They work together.

MCP connection vs custom MCP connector?

If a platform already has MCP, we connect from R1,000. If none exists, we build a bespoke connector from R3,500.

Does Buzz need Hermes?

A Buzz community can exist without one — but the agent is the value. Our Buzz packages include a Hermes-powered agent.

How long until it's working?

Basic setups typically complete the same day. Custom skills and connectors are scoped separately after we map your stack.

How do I start?

Email 8gents@agentmail.to with what you want — Hermes, Buzz, Local AI, or any combination. We reply with next steps and a clear quote.

What hardware do I need for a local model?

16GB RAM is the comfortable minimum. We audit your machine first and size the model to it — from 7B Q4 on a typical laptop up to 30B+ on bigger rigs. If your hardware won't run it, we tell you before you pay.

How good are local models vs ChatGPT?

Honest answer: strong for admin, drafting, code, and follow-ups — not frontier-grade for every task. The setup includes a guide to what runs best locally and when it's worth using a cloud model for one-off heavy lifts.

Will I really stop paying for credits?

Yes, for the workloads we install: no per-token bills, your data stays on your hardware, and the only running cost is electricity. You only pay for cloud models if you choose them for specific tasks.

Be the first to move

Your competitors are still answering email by hand.

Three ways to change that. Hermes agents from R2,500 once-off. Buzz communities from R2,500 once-off. Local AI engines from R7,500 once-off. Start simple, scale when your workflow demands it.

Start with a discovery call See pricing

Write us · 8gents@agentmail.to · Johannesburg · Worldwide

Your business,running on Hermes, Buzzand your own local AI.