AIHostBox
Client area Deploy an app
All 57 applications
CategoriesAI assistants & chatLLM runtimesAgents & RAGGateways & LLM opsVector & dataAutomationSpeech, image & videoSearch & web agentsSupport & CXML platformMCP servers
SolutionsPrivate team chatUnlimited automationCompany knowledge AIAI gateway & observability
More PricingGuides ComparisonsToolsPrivate LLM hosting LicensingSecurity SupportClient area

LiteLLM hosting

One OpenAI-compatible endpoint in front of every model provider you use.

Gateways & LLM opsMITfrom $5.99/mo
In short

LiteLLM hosting at AIHostBox means a dedicated, single-tenant LiteLLM instance on our infrastructure: we install it, pin it to a reviewed release, put it behind TLS on your domain, back it up nightly and keep it patched. Plans start at $5.99 per month and the admin account, the data and any API keys are yours.

100+ providersVirtual keysBudgetsFallbacksCachingSpend logs

LiteLLM is a proxy that speaks the OpenAI API and translates it to a hundred-odd providers behind it: OpenAI, Anthropic, Google, Bedrock, Azure, Groq, OpenRouter and any local server. Your applications keep one base URL and one key format while you change what is actually serving the request.

The operational value shows up quickly: virtual keys per team or per app, budgets and rate limits per key, fallbacks when a provider errors, caching, and a spend log that finally answers “what did we spend on AI last month, and on what”. Pair it with Langfuse and you have the observability half as well.

LiteLLM hosting plans

StarterPersonal use & small teams
2 GB RAM · 1 vCPU · NVMe
$5.99/mo
Order
PlusBusy workflows & communities
4 GB RAM · 2 vCPU · NVMe
$9.99/mo
Order
ProHeavy pipelines & many users
8 GB RAM · 3 vCPU · NVMe
$16.99/mo
Order

Every plan is a dedicated single-tenant instance with NVMe storage, nightly off-box backups, TLS and managed patching. Quarterly, semi-annual and annual terms are discounted at checkout. Model tokens are not included — you bring your own key.

How we run LiteLLM

The LiteLLM proxy container (docker.litellm.ai/berriai/litellm, main-stable) on port 4000 with Postgres 16 for keys, teams, budgets and spend logs, following the project’s docker-compose.

What we configure for you

  • LITELLM_MASTER_KEY generated per instance — the admin credential for API and UI
  • LITELLM_SALT_KEY generated once and stored safely, since it encrypts provider credentials in the database
  • DATABASE_URL on the bundled Postgres with STORE_MODEL_IN_DB=True, so models can be managed from /ui
  • UI_USERNAME and UI_PASSWORD for the Admin UI login
  • Your provider keys and model_list in config.yaml, or added through the UI
  • PROXY_BASE_URL plus GOOGLE_CLIENT_ID, MICROSOFT_CLIENT_ID or GENERIC_CLIENT_ID if you want single sign-on

Which plan fits

Starter (2 GB) is plenty for one team routing a handful of providers, since the proxy mostly forwards requests and waits. Plus (4 GB) suits several teams with spend logging on every call, and Pro (8 GB) is for heavy traffic or long request logs where Postgres grows quickly. Response latency is set by the upstream provider, not the plan.

The honest limits

SSO beyond five users and the other features LiteLLM marks as enterprise need a LiteLLM licence, which we do not supply. Logging full prompts and responses makes the database grow fast, so enable it deliberately.

What people run LiteLLM for

Central key custody

Provider keys live in one place; teams get revocable virtual keys.

Cost control

Hard budgets per key stop a runaway loop from becoming an invoice.

Provider failover

Route to a second provider automatically when the first degrades.

Local + cloud in one endpoint

Send cheap calls to your Ollama plan and hard ones to a frontier model.

Licence and version policy

ItemPosition
Upstream licenceMIT
Sold as managed hostingYes
TenancySingle tenant — one customer per instance
PatchingCritical advisories within 24 hours, high severity within 72

Our full position on licences, including the applications we decline to sell, is on the licensing page.

Reading on LiteLLM

ComparisonLiteLLM vs OpenRouterComparisonSelf-hosted Intercom alternativesComparisonLangfuse alternatives for LLM observabilityGuide · ArchitectureA self-hosted AI stack that actually holds togetherGuide · CostCutting an LLM bill without cutting qualityGuide · MCPRemote MCP server: how to host one your agents can reachGuide · Team chatPrivate ChatGPT for your team, on a server you control

Works well with LiteLLM

Langfuse logo

Langfuse

Tracing, evaluation and prompt management for LLM applications.

from $11.99/moDetails →Ollama logo

Ollama

Your private LLM server — pull a model, get an API, keep every token on your own box.

from $16.99/moDetails →Open WebUI logo

Open WebUI

A ChatGPT-style workspace for your team, pointed at any model you like.

from $5.99/moDetails →LibreChat logo

LibreChat

Multi-model chat with agents, plugins and conversation forking.

from $11.99/moDetails →

More Gateways & LLM ops we host

Opik logo

Opik

Open-source evaluation and monitoring for LLM applications.

from $11.99/moDetails →
Questions

LiteLLM hosting: your questions

Does LiteLLM store my prompts?
Only what you configure. Logging is yours to enable; the database lives on your instance, and we do not read it.
Can I put my Ollama plan behind it?
Yes — that is a common pairing. LiteLLM routes by model name, so a local model and a commercial one sit behind the same endpoint.
Where do my provider keys live, and how are they protected?
In LiteLLM’s Postgres database, encrypted with LITELLM_SALT_KEY when models are stored there. We generate that salt at setup and keep it with your instance secrets. LiteLLM’s docs are explicit that it cannot be rotated once models exist, because changing it makes the saved credentials unreadable, so it never changes during upgrades.
How do I log in to the Admin UI?
The UI is at /ui on your instance. You sign in with UI_USERNAME and UI_PASSWORD, or with the master key, then create further users. SSO through Google, Microsoft, Okta or any OIDC provider has been free for up to five users since v1.76.0; beyond that LiteLLM requires an enterprise licence.
Can I bring the config.yaml from my own LiteLLM deployment?
Yes. Upload your config.yaml and we mount it, so model_list entries and router settings keep working. To keep virtual keys, teams and spend history too, send a pg_dump of your LiteLLM database together with the same LITELLM_SALT_KEY; otherwise stored provider credentials will not decrypt.
How are LiteLLM upgrades handled?
LiteLLM releases often. We follow the main-stable image, run the database migrations before the new proxy version starts serving, and keep your config, virtual keys and spend logs in Postgres across versions. Release notes are worth a glance before each upgrade, since LiteLLM changes quickly.

Run LiteLLM without running a server

A dedicated LiteLLM instance, deployed and maintained by us. Month to month, cancel whenever.

Browse the catalogue Talk to a human