Practical overview
When it fits
Use LLM Gateway when an application needs a common OpenAI-compatible request interface across providers, or when a team wants consolidated usage and performance reporting. Its documented Docker route is relevant when the team prefers to deploy the gateway on its own infrastructure.
Who it is for
- Developers connecting applications to more than one language-model provider.
- Teams that want a shared interface for provider requests and API-key management.
- Operators who need usage, cost, latency and model-performance metrics.
Documented uses
- Route application requests to supported LLM providers through one API format.
- Manage provider API keys in one place.
- Review token use, request volume, costs and response times.
- Compare models using recorded performance and cost metrics.
- Choose between the documented hosted service and a self-hosted deployment.
Requirements & limits
The hosted service documentation says users create an account to obtain an API key. The project also documents self-hosting. Core functionality is described as AGPLv3-licensed; some enterprise features and multi-organization administration have separate licensing terms.
- The unified interface is compatible with the OpenAI API format.
- The README names OpenAI, Anthropic and Google Vertex AI among multiple provider integrations.
- Analytics cover requests, tokens, response times and costs; performance monitoring supports model comparisons.
- Self-hosting requires managing the deployment and following the documented Docker volume guidance.
- The repository describes enterprise licensing for commercial features and white-label licensing for multi-organization administration.
Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.
Deploy LLM Gateway with Docker-managed volumes
Start the documented unified Docker container with secret values and persistent data volumes.
Before you start
- Docker
- The LLM Gateway repository, including its unified-container script
- OpenSSL and the tr utility for generating secret values
Steps
Generate values for LLM_GATEWAY_SECRET and GATEWAY_API_KEY_HASH_SECRET using the documented OpenSSL commands.
Run ./scripts/run-unified-container.sh from the repository.
Expected outcome
The documented unified-container startup procedure is invoked with generated secrets; Docker-managed volumes are recommended for PostgreSQL data.
Guide limitations
- The documentation warns against directly bind-mounting a host directory to /var/lib/postgresql/data because PostgreSQL initialization may fail to set permissions.
- This guide covers the documented startup procedure, not application configuration or a tested deployment.
Send a chat-completions request to the gateway
Use the documented API request format to submit a chat-completions request.
Before you start
- An LLM Gateway API key stored as LLM_GATEWAY_API_KEY
- A model name and message content for the request
Steps
Send a POST request to https://api.llmgateway.io/v1/chat/completions.
Include Content-Type: application/json and Authorization: Bearer $LLM_GATEWAY_API_KEY headers.
Provide a model and a messages array in the JSON body.
Expected outcome
A chat-completions request is submitted through the documented gateway endpoint.
Guide limitations
- The example documents the request shape; it does not establish which model names are available to a particular account or deployment.
Sources & review
Source record details
- Source availability
- Active
- Sources checked
- 14 October 2026
- Original website title
- theopenco/llmgateway · Official homepage title · retained in its source language