An open-source AI gateway that combines more than 100 LLM providers into one OpenAI-compatible API and manages keys, budgets and costs
Information updated 2026-10-20, source: official site and public reports. Prices and benefits may change. Check the official site before you pay.
Pros
- Free open-source self-hosting
- Virtual keys/per-team budgets/rate limits
- Fallbacks/load balancing/cost-based routing
- Usage cost tracking
- Guardrails such as PII masking
Cons
- You must install and operate it yourself (no official hosting)
- SSO, audit logs and similar are paid Enterprise features
- Model usage fees are paid separately to each provider
These are opinions compiled by Pickx from public sources.
Cautions
In a PyPI supply-chain attack on 2026-03-24, litellm versions 1.82.7/1.82.8 contained credential-stealing malware and were pulled the same day (external security reports). If you ever installed those versions, rotate your keys. Pinning dependency versions is recommended
At a glance
LiteLLM is an open-source gateway (proxy) that unifies the LLM APIs of many companies into a single OpenAI-format API. It suits developers and platform teams who want to manage the keys/budgets/usage of several models in one place for a team or company. The software is free and you install it on your own server.
Main features/supported AI
| Feature/model | Description | Notes |
|---|---|---|
| Unified API | Call 140+ providers/1,800+ models through an OpenAI-compatible API | Python SDK/proxy server |
| Virtual keys/budgets | Issue keys per team/project, control model access, daily/monthly budget caps | Requests are blocked when the cap is reached |
| Routing | Load balancing, fallbacks, lowest-cost routing, caching | |
| Cost tracking | Aggregate usage cost per key/user/team/agent/MCP | Prometheus metrics, Langfuse logging integration |
| Guardrails | PII masking, prompt injection/secret detection | Integrations such as Presidio, Lakera |
Plans
| Plan | Price (monthly billing / yearly billing) | What you can do/limits | Features |
|---|---|---|---|
| Open Source (self-hosted) |
| 100+ providers, virtual keys, cost tracking, budget/rate limits, fallbacks, logging | Deploy with Docker/Helm/Terraform |
| Enterprise | Quote (annual) | SSO/SCIM, JWT authentication, audit logs, secret manager, multi-region, 24/7 SLA, air-gapped deployment | 30-day free trial key, can be purchased on AWS Marketplace |
Who it suits
- Open Source: Developers building a service that uses several LLMs who want to switch models without code changes
- Enterprise: Companies that must control all employees' AI API use together with SSO/audit logs
Compared with similar services
- LiteLLM vs OpenRouter: OpenRouter is a hosted intermediary you can use just by signing up (paid with credits), while LiteLLM is a self-hosted gateway you install yourself and connect each provider's keys to
- LiteLLM vs commercial AI gateways such as Portkey: LiteLLM's free open-source version has a wider feature range
- LiteLLM + Langfuse: Using the gateway (LiteLLM) together with tracing/evaluation (Langfuse) is a common setup
Notes
- Information checked: 2026-10-20
- LiteLLM official homepage
- LiteLLM official pricing
- Summary of the 2026-03 PyPI supply-chain attack (external source, Bastion)
Good for
5 views (3 today)
Prices converted at 1 USD = ₩1,329.49 (2026-10-19). Original price shown in brackets. Taxes as listed by each provider.
Comments
Sign in to leave a comment.
No comments yet. Be the first.
Your comments and nickname are public. If you see abuse or advertising, report it here and we will review and remove it.