gpt-load is a self-hosted gateway that lets applications call OpenAI, Anthropic, and Gemini APIs through one unified entry point. It centralizes API keys and subscription account management with request scheduling, failover, logs, and usage accounting.
Project overview
By using one mechanism for both API-key channels and subscription accounts, and letting clients keep their native OpenAI, Anthropic, or Gemini protocols, it removes the adapter and configuration sprawn typical of multi-provider AI setups.
Project type
Infrastructure
Deployment
Refer to project documentation
License
MIT
Best for
Developers and AI/operations teams who manage multiple AI provider channels, API keys, and subscription accounts and want unified scheduling, failover, logs, and usage accounting behind one self-hosted entry point.
Key capabilities
Routes client requests to official APIs, cloud platforms, model services, and compatible relays while clients keep OpenAI, Anthropic, or Gemini native interfaces.
Configurable weights, retries, cooldown, blacklisting, and session affinity reduce the impact of overloaded or failing credentials.
Connect SQLite, MySQL, or PostgreSQL via DATABASE_DSN.
gpt-load-windows-setup.exe installs a low-privilege Windows service with autostart and Start Menu shortcuts.
Limitations and risks
GPT-Load 2.0 cannot open, import, or migrate 1.x data in place; existing 1.x deployments cannot be upgraded in place.
Unknown client-supplied fields are not automatically forwarded; arbitrary parameter passthrough is not enabled.
The gateway proxies upstream model APIs; you must bring your own provider API keys or subscriptions, which may be paid services.
Getting started
Clone the repo, copy .env, and run docker compose up -d; verify http://127.0.0.1:3001/health; read the management key and sign in to the console; add a channel with API keys or OAuth subscription; create a group with models and policy; create an AccessKey and use it in the application. On Windows, gpt-load-windows-setup.exe automates service setup. No coding is required.
Evidence and sources
Release: v2.0.0-rc.37
GitHub project description: Self-hosted AI gateway for multi-channel, multi-credential setups — API keys and subscription accounts, scheduling, failover, request logs and usage.
README: If you are using 1.x, read [Moving from 1.x](#moving-from-1x) first. 2.0 cannot open, import, or migrate 1.x data in place.
README: One mechanism for API keys and subscriptions — Codex, Claude, Antigravity, Grok, and API-key channels share credential management, scheduling, and health handling.
README: Add a channel — Choose an upstream service and add one or more API keys. For subscription channels, complete the OAuth flow or import credentials as prompted.