This page explains how IntraLLM reaches your environment and what installation involves at a high level. The IntraLLM team performs or supervises installation as part of your engagement; this page sets expectations for the conversation with your IT and infrastructure teams.
Delivery paths
| Path | What is delivered |
|---|---|
| On-premise with hardware | Custom IntraLLM server(s) (2U, 4U, or MGX configurations), preloaded with the platform, delivered to your site. Your team completes network integration. |
| On your own hardware | IntraLLM software packages, installed onto compatible GPU infrastructure you already own. |
| IntraLLM Cloud | A hosted deployment provisioned by IntraLLM on dedicated machines; you receive access details and connect your identity provider. |
High-level requirements
Your infrastructure team should be ready to discuss:
- GPU capacity — sized by model count, model size, and expected concurrent users. Licences cover a GPU count (for example up to 4 or up to 8 GPUs per server).
- Networking — a reachable address inside your network for the platform, plus optional external access if you choose to use external model APIs or web search.
- Storage — capacity for models, documents, knowledge bases and generated media.
- Identity — details of your SSO/identity provider if you want Single Sign-On.
- Browsers — any modern browser on desktop or mobile; no client software is installed on user devices.
IntraLLM is delivered as a containerised stack and supports common enterprise deployment targets, including single-server installs and multi-node clusters. Exact instructions and commands are provided with your deployment package and by the IntraLLM support team.
Security note:
Installation credentials (initial administrator account, endpoints, and licensing) are sensitive. Store them in your password manager or secrets vault during handover, and change the default administrator password during initial setup.
What happens during installation
- Sizing and planning — the IntraLLM team confirms workloads (chat, documents, image/video generation, fine-tuning) and recommends a configuration.
- Provisioning — hardware arrives and is racked, or a cloud deployment is provisioned, or software is installed on your infrastructure.
- Platform bring-up — the IntraLLM services, model services and supporting components are started and verified.
- Licensing — your licence is applied to the deployment, enabling the features covered by your plan. See Licensing.
- Integration — network access, HTTPS, and optional SSO/identity provider connections are configured.
- Handover — administrator credentials and documentation are handed to your nominated admin team.
After installation
The first thing to do after installation is a short initial setup: create users and groups, connect model endpoints if needed, and import your first knowledge base. Follow the Initial Setup checklist, then continue to the Platform Walkthrough.
Note:
Need help at any point? Installation assistance is part of every support tier. Contact details are on the Contact Us page.