Startup
For founding teams and small businesses.
Coming soon
- 1 fine-tuned private model
- 3 AI agents (HR · Ops · Sales)
- Up to 25 seats
- Single-node deployment
- RAG over your documents
- Email support · 24h response
PRIVATE INTELLIGENCE PLATFORM
Your model. Your data. Your terms.
One private AI platform with nine specialised agents — fine-tuned on your knowledge, deployed on your infrastructure, with zero data sent to third parties.
POPULAR
Corporate
For scaling teams who need real depth.
Coming soon
- 3 fine-tuned private models
- All 9 AI agents included
- Up to 250 seats
- Multi-node, HA deployment
- Meeting intelligence + summariser
- Custom integrations
- Priority support · 4h response
Enterprise
For regulated industries and global units.
Coming soon
- Unlimited models & agents
- Unlimited seats
- On-prem · Air-gapped · Sovereign
- SSO · SAML · SCIM · audit logs
- Custom agent development
- SOC 2 · ISO 27001 alignment
- 24×7 dedicated engineering
COMPARE
Every plan, side by side.
Feature
Startup
Corporate
Enterprise
Private fine-tuned models
1
3
Unlimited
AI agents
3
9
9 + custom
Seats included
25
250
Unlimited
Deployment
Single-node
Multi-node HA
On-prem / Air-gap
Meeting intelligence
—
✓
✓
SSO · SAML · SCIM
—
✓
✓
Custom agent development
—
—
✓
Data sent to third parties
0%
0%
0%
HOW IT WORKS
Four steps to your private AI.
We learn your stack
We audit your data sources, workflows, and SOPs to scope the right agents.
Fine-tune on your data
Using Unsloth QLoRA on Llama 3.1, we build a model that speaks your business.
On your infrastructure
Containerised with Docker. Runs on your servers. Air-gapped if you need it.
FAQ
Questions, answered.
Does AfenixCore send any data outside our servers?
No. Zero outbound calls. The model, embeddings, and conversations all stay on your infrastructure — enforced at the network layer, not just by policy.
What hardware do we need?
A single GPU server handles Startup workloads. Corporate and Enterprise run on multi-GPU clusters we help you provision — on-prem or in your private cloud.
How long does deployment take?
Startup: 2–3 weeks. Corporate: 4–6 weeks. Enterprise: tailored to your compliance needs, typically 8–12 weeks.
Can we add custom agents?
Yes — included at Enterprise tier, available as add-ons for Corporate. The framework is built on CrewAI, so your engineers can extend it too.
What if we want to switch models later?
Because we use open weights, you can swap base models without rebuilding. Your data, embeddings, and fine-tuning datasets remain portable.
Let's build it.
We're preparing AfenixCore for launch. Join the waitlist to be first in line.