selfhost.directory

The index of self-hostable projects — ranked, categorized and served with live updates.
Own your data and stop giving it to big tech.

OpenGateLLM logo

Self-host OpenGateLLM

0.8.0

Open-source API gateway for self-hosted LLMs

BetaOpen SourceMIT
Updated11 hours ago
LanguagePython
Setup difficultyModerate
ThemeDark mode
Stars
218
Ports
8000 (API server), 8501 (Playground UI)
Requires
PostgreSQL, Redis
Docker
ghcr.io/etalab-ia/opengatellm/api
Install with Docker Compose
OpenGateLLM screenshot 1

What is OpenGateLLM?

OpenGateLLM is a freely licensed API gateway for generative AI, created for organizations that require complete authority over their data and infrastructure. It delivers a single OpenAI-compatible interface that directs requests to various model providers such as vLLM, Ollama, Mistral On-Prem, Text Embeddings Inference, and WhisperX. The system ships with built-in access governance covering roles, users, and organizations, along with usage controls including rate limiting, inference monitoring, model pricing budgets, and carbon footprint tracking. Both password-based login and OIDC-based Single Sign-On are supported. Deployment uses Docker Compose for quick starts or Kubernetes (Helm charts in progress), with PostgreSQL and Redis as required dependencies. The project comes from DINUM (French government digital service) and CentraleSupélec under the MIT license.

Should you self-host OpenGateLLM?

OpenGateLLM scores 46/100 on selfhost.directory (how scoring works) and ranks #24 of 39 in API Gateway. It has 218 GitHub stars.

It shipped 18 stable releases in the past year. The latest, 0.9.1, came out on 23 October 2026. New versions land roughly every 2 weeks.

Setup is moderate: expect some configuration beyond a single command. An official Docker image is available. It is MIT-licensed and written mostly in Python.

Its closest alternative on selfhost.directory is OmniRoute (88/100, 82 065 stars). It scores higher overall, so compare the two before you pick. See OpenGateLLM vs OmniRoute side by side.

Features

  • Access governance
  • Environmental footprint tracking
  • Load Balancing
  • Model aliases
  • Model pricing and budgets
  • OpenAI Compatible API
  • Playground UI
  • Rate Limiting
  • Single Sign-On (OIDC)
  • Unified model routing
  • Usage monitoring

OpenGateLLM README

How to install OpenGateLLM

Install OpenGateLLM with Docker Compose

Environment variables
CONFIG_FILE, COMPOSE_FILE, REFLEX_USE_GRANIAN, DEBUG, API_PORT, PLAYGROUND_PORT, POSTGRES_HOST, POSTGRES_PORT, POSTGRES_USER, POSTGRES_PASSWORD, REDIS_HOST, REDIS_PORT, REDIS_USER, REDIS_PASSWORD

git clone https://github.com/etalab-ia/OpenGateLLM.git cd OpenGateLLM pip install ".[dev]" make quickstart

Quickstart command launches API, Playground, PostgreSQL, and Redis in Docker containers. ARM64 hosts need DOCKER_DEFAULT_PLATFORM=linux/amd64.

How to configure OpenGateLLM

  • API_PORT8000Port for the API server

    Port for the API server

  • COMPOSE_FILEcompose.ymlDocker Compose file to use

    Docker Compose file to use

  • CONFIG_FILEconfig.ymlPath to the configuration file

    Path to the configuration file

  • DEBUGfalseEnable debug mode

    Enable debug mode

  • PLAYGROUND_PORT8501Port for the Playground UI

    Port for the Playground UI

  • POSTGRES_HOSTpostgresPostgreSQL host

    PostgreSQL host

  • POSTGRES_PASSWORDchangemePostgreSQL password

    PostgreSQL password

  • POSTGRES_PORT5432PostgreSQL port

    PostgreSQL port

  • POSTGRES_USERpostgresPostgreSQL user

    PostgreSQL user

  • REDIS_HOSTredisRedis host

    Redis host

  • REDIS_PASSWORDchangemeRedis password

    Redis password

  • REDIS_PORT6379Redis port

    Redis port

  • REDIS_USERredisRedis user

    Redis user

  • REFLEX_USE_GRANIANtrueUse Granian for Reflex backend

    Use Granian for Reflex backend

OpenGateLLM releases & version history

More in API Gateway

Rate this project

Click to rate — no account needed.

0 comments

Posting as guest
Loading comments…