Hermes AI What It Does Hosting Plans Hardware Pricing FAQ Join Founders Waitlist →
Australia-Wide · Remote & On-Site

AI That Saves You
Thousands
Per Month.

Cancel your ChatGPT, Claude, and Copilot subscriptions. We build AI agents that do 10× more work at 1/10th the cost — using smart routing across cheap and premium models so you only pay top dollar when it actually matters.

90% less
Token costs vs direct API
$199/mo
Hosted agents from
3–5 days
To go live
WhatsAppSlackTelegram DiscordEmailiMessage TeamsSignal
hermes-agent · neuralops-server · Brisbane SMB deploy
bashmemoryskills
Hermes: Online
llama3:8b + claude-sonnet
AU · UTC+10
Why It Costs So Little

Stop Paying
Top Dollar for Every Query.

Most businesses paying for ChatGPT Plus, Claude Pro, and Copilot are using a Ferrari to do the school run. We use smart routing — cheap models for routine work, premium models only when it actually matters. The result is dramatically lower running costs without sacrificing quality.

What you're doingThe naive wayThe NeuralOps wayYou save
Heavy admin (200 tasks/day)~$300/mo direct API~$9/mo via routing97%
Coding agent (8hrs/day)~$400/mo Claude Pro x2~$20/mo via routing95%
Research agent (heavy context)~$1,200/mo direct API~$91/mo via routing92%
Multi-tool SaaS stack~$280/mo subscriptions$199/mo all-inclusive29%

Estimates based on typical SMB workloads. Actual savings depend on volume and complexity. We'll calculate your specific numbers on the discovery call.

The trick most consultants won't tell you

Frontier models like Claude Sonnet cost up to 30× more than near-equivalent open models. For 90% of business tasks — drafting emails, summarising documents, scheduling, lookups — the cheaper models are genuinely indistinguishable. We route automatically: cheap models handle the bulk, premium models verify the hard stuff. You get top-tier quality at a fraction of the cost. Most consultants don't bother because it's harder to set up. We do, because it's the right answer.

Current Recommended Technology
🏛 Hermes Agent — by Nous Research · MIT Licence · 23,000+ GitHub Stars

Meet Hermes —
The Agent That Grows With You

Built by Nous Research — a serious AI research lab, not a solo developer — Hermes is currently our recommended foundation for most agent deployments. It self-improves, learns your business, and gets more capable the longer it runs.

Old way: Chatbots

Answers questions — can't take action on its own
Forgets everything between conversations
Locked to one channel — usually just your website
Needs someone to prompt it every single time
Gets dumber over time as your business changes
Can't connect to Xero, your CRM, or anything else

Hermes: Self-Improving Agent

Acts proactively — monitors, executes, reports without being asked
Persistent cross-session memory — knows your business forever
Lives on WhatsApp, Slack, Telegram, email — wherever you are
Built-in cron scheduling — 7am digest, weekly reports, auto follow-ups
Creates and improves its own skills from experience — gets smarter
Connects to 50+ apps: Xero, HubSpot, Gmail, calendar, Shopify...
Why Hermes is our current recommended foundation
→ Built by Nous Research — a real AI lab, not a community project
→ Self-improving loop: creates skills from experience automatically
→ 15+ platform gateways including WhatsApp, Slack, Home Assistant
→ 2,000+ community skills at agentskills.io ready to deploy
→ Runs on $5 VPS to GPU cluster — scales to any business size
→ MIT licensed — you own everything, zero vendor lock-in
Our Approach

Right Tool.
Right Job.

We don't sell one product. We assess your business, your budget, and your requirements — then deploy whatever technology actually solves your problem. The AI landscape moves fast. So do we.

Custom AI Chatbots
Website & Support Bots

Traditional chatbots still solve specific problems well — website FAQ bots, lead capture forms, product selectors. We build these when a chatbot is genuinely what's needed, not just what's trendy.

Web WidgetLead CaptureSupport
Claude API
Quality Intelligence Router

Anthropic's Claude is the best reasoning model available. We integrate it as a quality-checking layer — routing complex queries from cheaper local models to Claude only when needed.

Best ReasoningHybrid RoutingCost Optimised
Local AI Models
Private On-Premise Inference

Llama 3, Mistral, Phi-3 via Ollama — run entirely on your hardware. No data leaves your building. Zero per-query costs. The right choice for privacy-sensitive industries and high-volume workloads.

On-PremisePrivacy-FirstFree to Run
Flowise + RAG
Knowledge Base Builder

Upload your documents, manuals, and SOPs. We build a searchable AI knowledge base your agent draws from. Instant accurate answers from your own data — not generic AI guesses.

Document SearchRAG Pipelinepgvector
What Comes Next
Technology-Agnostic Promise

The AI landscape changes monthly. We stay current so you don't have to. If something better than Hermes ships tomorrow, we'll be evaluating it this week and deploying it next month.

Always CurrentNo Lock-InBest Available
The Real Question

"What exactly can
you get it to do
for our business?"

The most common question we get. Here's the honest answer — broken down into what it does automatically, what it does on demand, and what it always knows.

⏰ Runs Automatically — Without Being Asked
Every morning: summarises overnight emails, sends you a WhatsApp digest with priorities flagged
Scans your calendar and preps a briefing note for every meeting that day
Checks Xero/MYOB for overdue invoices and sends polite automated follow-ups
Monitors competitor websites — alerts you the moment their pricing changes
Watches your inbox 24/7 — new enquiries get a personalised response in seconds
Weekly sales report pulled from your CRM and texted to the owner every Monday
Stock level monitoring — alerts when anything drops below your threshold
Watches Google Reviews — flags negatives for immediate human attention
💬 On Demand — Just Type It in WhatsApp or Slack
"Find the Johnson contract and tell me what the payment terms were" — searched and returned in seconds
"Draft a quote for a 45m² deck, treated pine, supply and install" — formatted and ready to send
"Who called us last week that we haven't followed up?" — cross-referenced and listed
"Summarise this 40-page contract and flag anything unusual" — 1-page summary with risk flags
"Book a meeting with Sarah next Tuesday afternoon and send her an invite" — done
"What are our top 5 selling products this quarter?" — pulled from your data instantly
"Write a follow-up email to the client I met this morning" — drafted in your voice
"What do our terms of service say about refunds?" — exact answer from your documents
🧠 What It Always Knows (Your Knowledge Base)
Every product, service, price, and specification you offer
Your entire client history and what was agreed with each client
Every policy, procedure, and SOP your business operates under
Every quote, contract, and invoice ever issued
Your supplier list, lead times, and pricing agreements
Compliance and licensing requirements for your industry
Every staff member's role and current workload
Your competitors, their pricing, and how you compare
Industry-Specific Examples
Legal & Professional Services

Law Firm Agent

Reads every incoming brief overnight, drafts a summary, flags the deadline, posts to the relevant lawyer's Slack
Searches thousands of internal documents and precedents in seconds
Schedules consultations and logs billable time automatically
Data stays on-premises — never leaves your server
Self-HostedPrivacy-FirstSlack
Trades & Construction

Tradie Business Agent

Every quote enquiry on WhatsApp answered in 10 seconds — collects job details, auto-quotes standard jobs
Books into calendar, sends confirmations, follows up no-shows
Chases unpaid Xero invoices automatically with polite reminders
Answers product and pricing questions from your own database
WhatsAppXeroCalendar
Healthcare & Allied Health

Clinic Agent

Books and manages appointments via WhatsApp — confirms, reminds, follows up
Sends pre-appointment forms and post-visit care instructions
Answers patient FAQs on services, fees, and Medicare rebates
Fully self-hosted — patient data never leaves your premises
Privacy-FirstOn-PremiseWhatsApp
Retail & E-Commerce

Retail Agent

24/7 customer service on web chat, email, and social DMs
Order tracking, returns handling, product recommendations
Monitors stock, identifies slow-moving items, posts markdown recommendations
Escalates complex issues to a human with full context attached
Web ChatEmailShopify
Finance & Mortgage Broking

Finance Agent

Qualifies leads with the right questions before the advisor call
Sends rate change alerts and client portfolio digests automatically
Manages document collection — chases clients weekly until complete
On-premises deployment — AFS licence compliance ready
ComplianceSelf-HostedEmail
Property Management

Property Agent

Maintenance requests via WhatsApp — triages urgency, contacts tradesperson, confirms booking
Updates landlord automatically, closes the loop without PM touching it
Lease renewal reminders sent to tenants 90, 60, 30 days out
Inspection scheduling and confirmation handled end-to-end
WhatsAppEmailCalendar
Services

Everything We
Build and Deploy

01

Hermes AI Agent Setup

Our current recommended foundation for most deployments. Hermes configured and deployed on your channels, trained on your business data, connected to your apps and workflows. Self-improving out of the box.

HermesAll ChannelsSelf-Learns
02

Custom AI Chatbots

When a chatbot is the right tool — website FAQ bots, lead capture, customer service widgets. We build these properly, trained on your data, integrated with your CRM, and actually tested before handover.

Web WidgetLead CaptureClaude API
03

Knowledge Base & RAG

Your agent's memory — indexed from your documents, manuals, and SOPs. Staff and customers get instant accurate answers from your own data, not AI hallucinations.

FlowisepgvectorDocument AI
04

Workflow Automation

Connect your AI agent to Xero, HubSpot, Gmail, Calendly, Shopify and more. Automate the repetitive work that eats your team's time — quotes, invoices, bookings, follow-ups.

XeroHubSpotAny API
05

Self-Hosted AI Server

A dedicated AI server built, configured, and delivered to your premises. Runs Hermes and local AI models 24/7. Your data never leaves your building. Mandatory for legal, healthcare, and finance.

On-PremiseUbuntuOllama
06

Team Training & Workshops

On-site or remote workshops anywhere in Australia. Your staff leave knowing how to task, prompt, and get real value from your AI system — not just knowing it exists.

On-Site AUHands-OnRemote Option
How It Works

Live in 3–5 Days.

We assess your requirements first, recommend the right technology, then build and deploy it. No drawn-out projects — something working in your hands within the first week.

01

Discovery Call

30 minutes. We understand your business, your workflows, and what you actually need AI to do.

02

Tech Recommendation

We recommend the right stack for your situation — Hermes agent, chatbot, or hybrid. No upselling.

03

Fixed Quote

Clear, itemised proposal. You approve before any work begins. No surprises.

04

Build & Deploy

We configure, train, integrate, and test. Mostly remote — on-site where the project warrants it.

05

Handover

Full documentation and training. You own it — or we manage it on subscription.

AI Model Strategy

Hybrid Intelligence.
Near-Zero Running Cost.

We configure your AI to use the cheapest capable model for each task. Free local models handle the bulk of the work. Claude handles the hard stuff. You get great quality at a fraction of the cost.

☁️

Cloud API Only

Claude · OpenAI · xAI Grok

Best quality on every query. No hardware. Ideal for low-volume or quality-critical cases where per-query cost is acceptable.

 Zero upfront hardware cost
 Best reasoning quality available
~ Ongoing per-token costs
~ Data sent to third-party servers

Hybrid Stack

Local model triage + Claude for hard tasks

Free local model handles 90% of queries. Claude Sonnet handles complex reasoning automatically. ~$0.80–2/month API cost per business. This is what we run on our hosted infrastructure.

 Near-zero monthly AI costs
 Claude quality when it counts
 Fully automatic routing
 Best value for most businesses
🖥️

Fully Local

Ollama · Llama 3 · Mistral · On-premise only

100% on your hardware. No data ever leaves your building. Zero ongoing AI costs after setup. Required for legal, healthcare, finance, and defence.

 Data never leaves premises
 Zero ongoing AI costs
 AU Privacy Act compliant
~ Hardware investment required
Managed Hosting

We Run It.
You Just Use It.

Your AI agent hosted on our Australian infrastructure. Hybrid local + Claude routing. Flat monthly fee — setup cost waived. Works anywhere in the world. Cancel anytime after minimum term.

Starter Agent
$199/mo
Single channel. Up to 500 queries/month. Sole traders and small business.
  • Hermes agent on 1 channel
  • Knowledge base up to 200 pages
  • Hybrid local + Claude quality check
  • 3 custom skills included
  • Monthly performance report
  • Email support
  • Multi-channel deployment
  • CRM / app integrations
Get Started
Business Agent
$399/mo
Multi-channel. Up to 2,000 queries/month. Professional services, clinics, agencies.
  • Hermes on 3 channels (your choice)
  • Knowledge base up to 1,000 pages
  • Hybrid local + Claude quality check
  • 8 custom skills included
  • CRM / Xero / calendar integration
  • Proactive scheduled tasks (cron)
  • Priority support — next business day
Get Started
Enterprise Agent
$799/mo
Unlimited channels. Multiple agents. Full integrations. Medium business, multi-location.
  • Multiple Hermes agents
  • Unlimited channels & queries
  • Unlimited custom skills
  • Full app ecosystem integration
  • Monthly strategy review call
  • 4-hour support SLA
  • Dedicated account management
Get Started

All prices ex-GST. 3-month minimum term. Setup fee waived (normally $2,800). Data stored in Australia. International clients welcome, billed in AUD.

Hosted vs Self-Hosted — Which Is Right?
FactorHosted SubscriptionYour Own Hardware
Upfront cost$0 (included in plan)Est. from $6,000 AUD
Monthly ongoing$199 – $799Electricity only (~$65–95)
Data locationAustralia (encrypted, isolated)Your building only
Uptime managementNeuralOps manages itYou (retainer available)
Model updatesAutomatic, includedManual or via retainer
Works globallyAny country, any timezoneOn-site setup required
Best forSMBs wanting simplicityPrivacy-critical, high-volume
Break-even vs hosted~12–18 months (then free)
Own Your Hardware

Your AI Server.
Your Premises.

We build, configure, and ship AI servers anywhere in Australia. Runs Hermes and local AI models 24/7 on-site. Your data never leaves. One-time cost — no ongoing AI fees.

Note on pricing: Prices below are guide estimates in AUD only. Component costs fluctuate with market availability, GPU generation, and the quality of parts selected. Your final quote reflects current market pricing and your exact requirements. We always present options at multiple price points.
Starter

API Node

Est. from (AUD)
$6,000
Mini PC · 16–32GB RAM · 1TB SSD
Connects to Claude / OpenAI / xAI APIs
Guide estimate — exact spec quoted per build
  • Hermes fully configured
  • 10+ channels connected
  • Open WebUI staff interface
  • Cloudflare tunnel — secure remote access
  • Ubuntu Server, fully set up
  • Ships anywhere in Australia
Ongoing: API usage only (~$30–150/mo)
Best Value
Professional

Local AI Server

Est. from (AUD)
$12,000
Custom Build · RTX-class GPU 16GB+
Ryzen/Intel · 32GB RAM · 2TB NVMe · UPS
Guide estimate — spec and cost vary by GPU availability
  • Hermes + local AI models running
  • Llama 3, Mistral, Phi-3 preloaded
  • Hybrid routing to Claude configured
  • 5–10 concurrent users
  • Automated nightly backups
  • UPS power protection included
Ongoing: Electricity only (~$65–80/mo)
Enterprise

Pro AI Workstation

Est. from (AUD)
$18,000
Custom Build · High-VRAM GPU 24GB+
High-core CPU · 64GB RAM · 4TB NVMe
Guide estimate — varies significantly by component choice
  • Runs 70B+ parameter models locally
  • Multiple Hermes agents simultaneously
  • 15–25 concurrent users
  • Enterprise RAG knowledge base
  • Full network integration
  • On-site setup in major cities
Ongoing: Electricity only (~$85–110/mo)
Custom

Dual GPU Server

Price on application
POA
Scoped entirely to your requirements
Dual GPU, rack or tower options
Full quote provided after discovery call
  • 50+ concurrent users
  • Multiple large models simultaneously
  • Rack or tower form factor
  • Full network & AD integration
  • Managed maintenance included
  • National on-site support
Always included:  Ubuntu Server · Hermes Agent · Ollama · Open WebUI · Flowise · Nginx · Cloudflare tunnel · Tailscale remote access · Automated backups · Full documentation · 2 weeks post-delivery support. All prices are guide estimates in AUD, ex-GST, subject to current market pricing. Shipped Australia-wide. On-site setup: Sydney · Melbourne · Brisbane · Adelaide · Perth.
Consulting Rates

Transparent
Flat Rates

All work performed personally — no offshore handoffs, no junior staff. Remote by default, on-site available anywhere in Australia.

Hourly Rate
$350 AUD/hr
Pay-as-you-go. 2hr minimum. Prepaid 10hr block: $3,000 (save $500).
  • Remote anywhere in the world
  • On-site AU major cities
  • All configs and code included
Book Hours
MOST POPULAR
Starter Project
$2,800 AUD
8 hours. One defined deployment — assessed, configured, connected, tested, handed over.
  • Discovery + tech recommendation
  • Full agent setup
  • Staff training (1hr)
  • 2 weeks post-launch support
  • All code and configs included
Get Started
Business Project
$6,500 AUD
20 hours. Multi-agent or multi-system deployment, full knowledge base, team workshop.
  • Multiple agents or systems
  • Full knowledge base (RAG)
  • Half-day team workshop
  • 6 weeks ongoing support
  • Docs & runbooks
Enquire

All prices ex-GST. Travel quoted separately for on-site outside SA. Maintenance retainers from $300/month.

Where We Work

Australia-Wide.
Worldwide.

Sydney

Finance, legal, tech, retail. On-site available.

Melbourne

Healthcare, professional services, hospitality. On-site available.

Brisbane

Construction, trades, retail. On-site by arrangement.

Adelaide

SA home base. Full on-site. Fastest turnaround.

Perth

Resources, mining, professional services. Remote + shipped.

Worldwide

Hosted plans work anywhere. International clients welcome.

About NeuralOps

Based in Adelaide.
No Dogma. Just Results.

NeuralOps is an Australian AI consulting practice. We build AI agents, chatbots, and automation systems for businesses across Australia and internationally — using whatever technology actually solves your problem, not whatever we happen to sell.

Right now that usually means Hermes Agent as the foundation, Claude for quality routing, and local models for cost efficiency. Next month it might be something else. We stay current so you don't have to.

Every engagement is handled personally. No account managers, no offshore teams. Direct, accountable, fast. Remote-first — on-site anywhere in Australia when the project warrants it.

Current Preferred Stack
Hermes Agent 🏛Agent Framework
Claude (Anthropic)Quality Router
Llama 3 / MistralLocal Inference
OllamaModel Runner
Open WebUIStaff Interface
FlowiseRAG / Knowledge
pgvector / ChromaDBVector Store
Ubuntu ServerOS
Cloudflare + TailscaleNetworking
FAQ

Common
Questions

Hermes is currently the best open-source agent framework available. It's built by Nous Research — a serious AI research lab, not a community project — has 23,000+ GitHub stars, supports 15+ messaging platforms, includes a self-improving skills system with 2,000+ community skills, and is MIT licensed so you own everything. It also runs on anything from a $5 VPS to a GPU cluster. That said, we recommend it because it's the best current tool — not because we're locked into it. If something better ships, we'll evaluate and adopt it.

We build those too. A Hermes agent is overkill for some use cases — if you need a website FAQ bot that answers common questions and captures leads, a simpler chatbot built on Claude API is the right tool and we'll tell you that. We assess your actual needs and recommend accordingly. Our job is to solve your problem, not to deploy the most complex thing possible.

The short answer: anything that involves reading information, writing responses, searching documents, sending messages, booking appointments, or triggering actions in connected software. Practically speaking — answering enquiries 24/7, drafting quotes and emails, chasing unpaid invoices, summarising overnight emails, booking appointments, searching your internal documents in seconds, and running scheduled reports while you sleep. The "What It Does" section above has 24+ specific examples with real industry use cases.

On our hosted subscription, your monthly fee covers everything. For self-hosted hardware with local models, your ongoing cost is electricity — roughly $65–95/month for typical business workloads. Our hybrid routing sends around 10% of queries to Claude Sonnet, which typically costs $0.80–$2.00/month per business. The economics are dramatically better than any SaaS AI subscription.

Hosted clients: all infrastructure is in Australia, data is encrypted at rest, each client's data is completely isolated. Nightly encrypted backups. We operate under the Australian Privacy Act 1988. For self-hosted clients: your data never leaves your premises — not to us, not to any cloud. For industries with strict data sovereignty requirements (legal, healthcare, finance, defence), self-hosted is the only appropriate option and we'll tell you that upfront.

No — the prices shown are guide estimates in AUD. Component costs fluctuate with market availability, GPU generation, and the specific quality of parts selected. GPU pricing in particular varies significantly. We provide an exact quote after a discovery call where we understand your budget, user count, and requirements. We always present options at multiple price points.

Yes — for hardware installs, complex enterprise setups, and team workshops in Sydney, Melbourne, Brisbane, and Perth. Travel costs are quoted transparently upfront. Most deployments are handled entirely remotely with identical results — the agent runs on your infrastructure either way.

Yes. Hosted subscriptions work anywhere in the world — setup is remote, billing in AUD, data stored in Australia. Hardware can be shipped internationally on request. We work across all timezones.

Get In Touch

Book a Free
Discovery Call

30 minutes, no cost, no obligation. We assess your business requirements, recommend the right technology honestly, and tell you what's realistic. If a simple chatbot is what you need, we'll say so — not upsell you to something bigger.

Based
ServingSydney · Melbourne · Brisbane · Adelaide · Perth · Worldwide
Email
Phone
HoursMon–Fri 8am–6pm ACST (UTC+9:30)
Quick Start Options
Hosted Agent — Live in 3 days, from $199/mo, works anywhere
Hardware Package — Est. from $6,000 AUD, shipped AU-wide
Hourly Consult — 2hrs remote, we map what AI can do for you