LiteLLM: review, pricing, security and alternatives

What LiteLLM does

LiteLLM belongs to the agent infrastructure family: technical building blocks for agents: memory, execution sandboxes, model gateways, observability, evaluation. On our grid it ranks #13 of the 28 projects of this family listed in the directory.

It is not an agent in itself: it is a component other agents use (memory, isolated execution, web access, model gateway or monitoring). You pick it based on the agent or framework it has to equip.

You use it from: API, developer SDK, web app, and terminal (CLI). Usable models: Claude (Anthropic), GPT (OpenAI), Gemini (Google), DeepSeek, Qwen (Alibaba), Mistral, Llama (Meta), Kimi (Moonshot), GLM (Zhipu), Grok (xAI), local models (Ollama, llama.cpp, vLLM…), OpenRouter, and any OpenAI-compatible model. It can run on your own machine or server.

Who is LiteLLM for?

It is for teams running agents in production that need a reliable component: isolated execution, memory, model gateway or observability. Typical tasks people give it:

• isolate code generated by an agent
• track model call costs and errors
• give an agent long-term memory

It is a good candidate if you are comfortable in a terminal, your data must stay on your machines, and you want the option of a local model, with no API fees.

Access and human control

Its documentation does not detail the access granted to the agent: check the settings before connecting it to your accounts.

We could not establish whether it asks for human approval before sensitive actions: test it first on an unimportant account or folder.

LiteLLM pricing

The code is public but the repository declares no licence: reuse is not formally allowed. The real cost depends on hosting and, where relevant, the vendor's hosted offer.

Getting started with LiteLLM

1. Install it: the project ships as a Python package (install with pip or uv). The exact command is in the repository README.
2. Start with the minimal example from the docs, then add one tool at a time and check what the agent does with it.
3. Give it a short, checkable task, review the result and its action log, then raise the difficulty step by step.

Security: risk level

An infrastructure component sees prompts, keys and data pass through: self-host it or read the service's retention policy.

Points found for LiteLLM:

• No licence: legally shaky for company use.

Three habits: least privilege (minimal rights, dedicated keys), human approval for irreversible actions, a sandbox for code execution. The AI agent security check reviews your setup in a few minutes.

Project activity and releases

Created on July 27, 2023, the BerriAI/litellm repository has 61,965 stars, 12,636 forks and 1,820 open issues, with 55,118 commits in total. It received changes this month.

Over the last 52 weeks: 34,459 commits, an average of 662.7 per week.

1,530 releases in total; 18 in the last twelve months. Latest release: v1.106.3, on October 24, 2026.

LiteLLM vs other agent infrastructure

GitHub stars of the best-rated open-source projects in the agent infrastructure family. Stars measure attention, not quality: a heavily promoted project collects them fast, a specialised tool will have few.

LiteLLM and other agent infrastructure at a glance

The 7 best-ranked agents in the same family, 7 of them open source. The score out of 100 only applies to open-source projects; for proprietary services we show the published entry price. A dash marks information the documentation does not state.

How we rated it

Score: 79/100, computed only from public, verifiable data, with no subjective testing or partnership:

• Adoption 34/40: 61,965 stars on a log scale (40 points from about 200,000).
• Activity 30/30: last commit less than a month ago, 18 releases in six months, 34,459 commits in a year.
• Framework 15/30: licence, default human approval, self-hosting, model choice and documentation site.

The score does not say whether the agent will succeed at your tasks: it measures how solid the project is. The full grid is on the method page.