Model infrastructure
LiteLLM Gateway
Private model gateway with spend controls.
Automated deployment — typically ready within minutes
Overview
What it is.
A managed LiteLLM proxy that gives your applications one OpenAI-compatible endpoint across providers — with rate limits, key management, spend tracking, and fallbacks, operated by QIOI.
Best for
- Developers
- Teams using multiple model providers
- Agencies managing client usage
Scope
Exactly where QIOI's responsibility ends.
A managed service is only trustworthy when its boundaries are explicit. This is the boundary.
What QIOI manages
- Dedicated server resources on a tested plan
- Automated application deployment from a pinned template
- SSL certificate and qioi.app subdomain
- Health monitoring with automatic restarts
- Scheduled encrypted backups
- Managed upgrade path with rollback support
- QIOI dashboard: logs, metrics, backups, domains
- Virtual key management and per-key budgets
- Provider fallback and retry configuration
Not included
- —Third-party model API usage (billed by your model provider)
- —Custom application or workflow development
- —Data preparation beyond package limits
- —Unlimited technical consulting
Resource plans
Recommended capacity.
LiteLLM Gateway requires at least 4 GB of RAM. Plans include hosting and management in one line item.
Plan S
$39/mo
- vCPU
- 2
- RAM
- 4 GB
- Storage
- 80 GB
- Transfer
- 4 TB
- Backups
- 40 GB
Single-user workspaces and light workflows
Plan M
$69/mo
- vCPU
- 4
- RAM
- 8 GB
- Storage
- 160 GB
- Transfer
- 8 TB
- Backups
- 80 GB
Active business use, small teams
Deployment configuration
What a deployment includes.
The deploy flow confirms these options before provisioning. Nothing is hidden behind defaults.
- Plan
- Recommended: S / M
- Region
- ca-central, us-east, eu-central
- Domain
- qioi.app subdomain + custom domain
- Backup frequency
- Daily (Operate tier and above)
- Update channels
- stable / lts
- Model provider
- External model API usage is billed by your provider, not QIOI
Preview
How a LiteLLM Gateway deployment is structured.
Conceptual preview — real interface screenshots ship with the product
Your users
Browser / API clients
Managed LiteLLM Gateway
Monitored · backed up · updated
QIOI operations
Health checks · rollback · alerts
FAQ
LiteLLM Gateway questions.
Who holds the provider API keys?+
You do. Keys are stored encrypted on your instance and never logged in plaintext. You can rotate or revoke them at any time from the dashboard.