Private AI Deployments
What a private Emerge.ai deployment is, and who each of the three — Leona for enterprise, Nala for the edge, Simba for the desktop — is built for.
Emerge.ai is normally used as a hosted service: you sign in, choose a model, and Emerge.ai handles the request. A private deployment is the alternative — Emerge.ai running in an environment your organization controls instead.
Emerge.ai publishes three of them, and what separates them is where each one runs and who it is built for: Leona for enterprises, Nala for the edge and on-premises, and Simba for the desktop.
A private deployment is a different thing from the Emerge Private Hosted group in the model picker. That group is models Emerge.ai hosts itself, offered inside the ordinary service — see Models & Auto-Selector. A private deployment moves Emerge.ai itself into your own environment.
Leona — for enterprise
Leona is the largest of the three, built for enterprises that need strict privacy and substantial computing power. It sits between your existing systems and the AI models, giving every one of them the same interface, and it runs in cloud environments and data centres.
Leona is the one built for enterprise scale; the other two are for a single site or a single machine.
Nala — for the edge and on-premises
Nala runs AI at your own locations — branch offices and on-premises sites. Emerge.ai's own page for it leads with "Secure, Autonomous, Hybrid AI Operations at your location": it runs where your work is, while your security and compliance controls stay in place.
Two things it is built around:
- Access and identity controls. Audit logs record user actions, system events, and data access. SCIM support keeps user accounts in step with the identity system you already run.
- Large working context. Nala handles high-volume work in a single session — hundreds of transcripts, dozens of documents over 100 pages, or 100,000 lines of code.
Nala is optimized for Cisco Unified Edge, Dell PowerEdge, and HPE ProLiant.
It also connects to the tools a team already uses. Connectors covers what a connector is in Emerge.ai and how you choose which of its tools Emerge.ai may use; which connectors a given deployment offers is a question for Emerge.ai.
Simba — for the desktop
Simba runs on one machine. Emerge.ai builds it for smaller businesses, executives, and power users who put heavy, confidential work through AI and want it kept on hardware they control.
Simba performs best on Apple Mac Studio hardware; Emerge.ai publishes the recommended configuration alongside the deployment's setup requirements.
Getting one
Private deployments are arranged with Emerge.ai directly. Contact Emerge.ai to start that conversation.
Definitions
- Private deployment — Emerge.ai running in an environment your organization controls, rather than as a service Emerge.ai hosts for you.
- Leona — the enterprise deployment, running in cloud environments and data centres. Sits between your systems and the AI models and gives them all one interface.
- Nala — the edge deployment, for branch-office and on-premises sites.
- Simba — the desktop deployment, running on one machine.
- Emerge Private Hosted — a group in the model picker holding models Emerge.ai hosts itself. Part of the ordinary hosted service, not a private deployment.
Which deployment fits
| Deployment | Where it runs | Built for |
|---|---|---|
| Leona | Cloud environments and data centres | Enterprises needing strict privacy at scale |
| Nala | Your own sites — branch offices and on-premises | Organizations carrying security and compliance requirements at each location |
| Simba | A single desktop workstation | Smaller businesses, executives, and power users with confidential, heavy workloads |
Related topics
- Models & Auto-Selector — the groups in the model picker, including Emerge.ai's own hosted models
- Lockdown Mode — restricting AI processing to the models Emerge.ai has approved, inside the ordinary service
- Plans, Tokens & Billing — what the published plans include
- Troubleshooting
Model selector
The models in each group of the model picker as last checked against the product, with the picker itself named as the live source because Emerge.ai reconfigures the list over time.
Release Notes
Every published Emerge.ai release, newest first, with a summary of what shipped in each and a link to the full notes.
