LiteLLM
One OpenAI-style gateway to 100+ model providers, with spend controls
Overview
LiteLLM is an open-source AI gateway and Python SDK that routes calls to 100+ LLM providers through one OpenAI-style API, with virtual keys, budgets, rate limits, fallbacks, logging and MCP and A2A gateways, now on a Rust core (beta). Free to self-host (MIT); Enterprise priced by request volume.
Pricing
Pricing shown for reference only. These figures reflect RECATOOLS research as of 1 Oct 2026 and may be out of date or incomplete. This is not financial or purchasing advice — always confirm the current price on the provider’s official website before making any decision.
- 100+ providers, one OpenAI API
- Virtual keys, users and teams
- Spend tracking, budgets and rate limits
- Fallbacks, logging and Prometheus metrics
- SSO, SCIM and audit logs
- Secret managers and key rotation
- Multi-region control plane
- 24/7 SLAs available, self-hosted or air-gapped
What you can produce with LiteLLM
- OpenAI-compatible gateway to 100+ LLM providers
- Virtual keys, budgets and rate limits
- Model fallbacks and load balancing
- Spend tracking, logging and Prometheus metrics
- MCP and A2A agent gateways
- Python SDK and liteagents SDK
LiteLLM is the most widely used open-source way to put many model providers behind one API with cost controls, and its move to a Rust core should cut overhead. Its security record in 2026 is the main caution: a poisoned PyPI release and three actively exploited vulnerabilities, so anyone running it must patch quickly and lock down the admin interface. The open-source version is fully usable; Enterprise adds SSO, audit logs and support on quoted annual contracts. A sensible choice for teams that will maintain it carefully.
What we found
LiteLLM is an open-source AI gateway and Python SDK that lets applications call more than 100 model providers — OpenAI, Anthropic, Google, AWS Bedrock, Azure, local models and many others — through one OpenAI-style API. As a proxy server it adds virtual keys for users and teams, spend tracking, budgets and rate limits, fallbacks between models, request logging and Prometheus metrics, and it now also acts as a gateway for MCP tools and agent-to-agent traffic. In June 2026 the gateway began moving to a Rust core while keeping the same configuration; the Rust version is labelled beta, and the company claims sub-millisecond overhead, by its own benchmark. A new liteagents SDK routes each agent turn to a suitable model, and releases now arrive almost daily across four supported minor versions.
Security has been a recurring concern. In March 2026 two compromised versions were published to PyPI as part of a wider supply-chain campaign; an SQL-injection flaw, CVE-2026-42208, was exploited within 36 hours of disclosure; and by September 2026 three LiteLLM vulnerabilities were on the US CISA list of known exploited vulnerabilities, with further critical advisories at the end of September. Teams running it should update promptly and restrict access to the admin interface.
LiteLLM is developed by BerriAI, legally Berrie AI Incorporated, founded in San Francisco in 2023 by Krrish Dholakia and Ishaan Jaffer, which went through Y Combinator's Winter 2023 batch and raised a $1.6 million seed round. Its site quotes NVIDIA, Netflix, Okta, Lemonade and AT&T as users, and it holds SOC 2 Type II certification.
The open-source gateway is free to self-host under the MIT licence, except for its enterprise directory. LiteLLM Enterprise, sold on annual contracts sized by request volume and through the AWS and Azure marketplaces, adds SSO and SCIM, audit logs, secret managers, multi-region control and dedicated support, with 24/7 SLAs available; no price is published, and a 30-day trial key needs no card.
It suits engineering teams wanting one controlled entry point to many model providers.
Compiled by RECATOOLS from public sources: vendor documentation and pricing, code repositories and published reporting.
About this listing
This entry was compiled from publicly available data including LiteLLM's official website, press releases, documentation, and reputable third-party publications. RECATOOLS is not affiliated with LiteLLM unless explicitly stated.
Third-party AI tools update their pricing, features, availability, and policies frequently. Information here may be outdated by the time you read this — we make reasonable efforts to keep listings current, but cannot guarantee absolute accuracy.
For the latest details, please refer to LiteLLM directly →
Spotted something out of date? Suggest an update →
From our guides
Guide · 23 min read
Serve an LLM with vLLM on CPU in Docker: Learn the Production Engine Without a GPU (2026)
We ran vLLM v0.25.1's official CPU image in Docker on Apple Silicon, dodged a 10...
Guide · 14 min read
Self-Host LiteLLM with Docker: One OpenAI-Compatible Endpoint for Every Model (2026)
We ran LiteLLM v1.93.0 against a local Ollama model in Docker: pinned tags, a 4-...
Alternatives to LiteLLM
More in Code & Dev Tools
