OSS Tanbou
View on GitHubProject website

unify 100+ LLM providers behind OpenAI-compatible interfaces and centralize routing, cost controls, guardrails, and logging in an AI gateway

OSS scale score 322.8OSS health 96Update pace 6,441 commits/30d
About these scores

OSS scale score is an unbounded metric that log-compresses and weights Stars, Watchers, Forks, and Contributors. Discovery score is the current OSS scale score minus the score at discovery. Update pace is commits in the last 30 days, growth momentum is the OSS scale score difference within the recent observation window, and OSS health is a 0–100 rating based on available recency, Community Health, and release data.

Stars
60,090
Primary language
Python
License
Not determined
Repository last updated
Oct 3, 2026
On this page

Overview

LiteLLM is a Python SDK and AI gateway that presents a unified interface for more than 100 LLM providers including OpenAI, Anthropic, Gemini, Bedrock, and Azure. Its proxy can centralize virtual keys, spend tracking, guardrails, load balancing, and logging.

Features and best fit

Based on official documentation; not hands-on tested · Content checked:

Key features

Call more than 100 LLM providers through common interfaces

The Python SDK and proxy reduce provider-specific SDK and request-format differences through OpenAI-compatible and native endpoints.

Sources: [1]

Centralize routing, load balancing, and spend controls

A shared gateway can manage virtual keys, spend tracking, routing, load balancing, guardrails, and logging in one control layer.

Sources: [1]

Extend the gateway to MCP tools and agents

The project also documents MCP gateway and A2A agent integrations, allowing model and tool traffic to share a gateway boundary.

Sources: [1]

Best fit

Fits teams isolating provider choice and multi-provider policy from application code

It is useful when authentication, endpoints, fallbacks, and usage policy should be managed centrally instead of repeated across applications.

Sources: [1]

Before adoption

Protect the secrets, prompts, and usage data concentrated at the gateway

Centralizing model traffic also centralizes credentials and potentially sensitive request metadata, so network boundaries, key management, log access, and retention require explicit design.

Sources: [1]

Do not flatten the entire repository to a single MIT license

The v1.103.2 LICENSE assigns separate terms to the enterprise/ directory and MIT to content outside it. Because repository metadata reports NOASSERTION, the repository-level SPDX is left unset and component terms should be checked directly.

Sources: [3]

Pin releases and regression-test provider compatibility

As of October 3, 2026, the latest stable release is v1.103.2. Provider APIs can also change independently, so upgrades should regression-test requests, streaming, errors, and cost accounting for the providers in use.

Sources: [2][1]

Official sources

  1. [1]LiteLLM v1.103.2 README(2026-10-03)
  2. [2]LiteLLM v1.103.2 release(2026-10-03)
  3. [3]LiteLLM v1.103.2 licensing(2026-10-03)
Supplemental curator note

LiteLLM is useful when applications need one interface across many model providers. Because the gateway can centralize API keys, prompts, and usage metadata, authentication, secret storage, log retention, and provider-specific data policies should be designed before production use.

Try it in 3 steps

  1. 1

    Install LiteLLM 1.103.2 in an isolated environment

    Pin the stable release locally without configuring any provider API keys.

    python3 -m venv .venv && . .venv/bin/activate && python -m pip install litellm==1.103.2
  2. 2

    Verify the installed version

    Check only local package metadata without sending requests to an external model provider.

    python -c "import importlib.metadata; print(importlib.metadata.version('litellm'))"
  3. 3

    Import the shared completion interface

    Confirm the primary SDK interface loads successfully without provider credentials.

    python -c "from litellm import completion; print(callable(completion))"
Check the official README

Growth

Growth trends · Last 30 days

60,090 Stars

Trend data is still being collected.

Development activity

Last 90 days · weekly

Commits (last 30 days)
6,441
Open PRs
3,796

Development activity is still being collected.

Built with

Categories and tags

Categories

Tags

GitHub data

GitHub dataView detailed GitHub data
View on GitHub

GitHub Topics

  • anthropic
  • langchain
  • llm
  • llmops
  • openai
  • ai-gateway
  • azure-openai
  • bedrock
  • gateway
  • openai-proxy
  • vertex-ai
  • llm-gateway
Stars
60,090
Forks
11,999
Watchers
232
Open issues
1,799
Contributors
374
Owner type
Organization
Primary language
Python
License
Not determined
Repository last updated
Oct 3, 2026
Write a related article

Share a guide or use case for this OSS in Markdown. Articles are published after administrator approval.

Report incorrect information

Tell us if any listing information is incorrect or outdated.

After reading this page, do you know what to do next?