What is GPT-Load used for?

GPT-Load is a self-hosted AI gateway that puts many model channels, API keys and subscription accounts behind a single endpoint with scheduling, failover, request logs and usage accounting. It is a good fit for pooling several AI provider keys, routing coding agents through one gateway and tracking AI usage per credential.

Is GPT-Load open source?

Yes. GPT-Load is open source under the MIT licence. Its source code is on GitHub at tbphp/gpt-load and is written mainly in Go.

Yes. GPT-Load is open source, so the software itself is free to use.

Can I self-host GPT-Load?

Yes. GPT-Load can be self-hosted on your own server or infrastructure.

What are some alternatives to GPT-Load?

Similar open-source tools in the AI Infrastructure category include CoAI, New API and LLM Gateway. SaaS products in the same category include AllToken, Costbase and AI Stats.

Is GPT-Load actively maintained?

Yes. The most recent commit to GPT-Load was on 2 October 2026, and the latest release is v2.0.0-rc.41, published on 2 October 2026. The project has 7k stars on GitHub.

GPT-Load: open-source AI infrastructure

About GPT-Load

GPT-Load is a self-hosted AI gateway for setups that use several model channels and credentials. It puts API keys, subscription accounts, traffic scheduling, error handling, request logging and usage accounting behind a single entry point, so applications talk to a single endpoint instead of juggling providers.

Its topics show support for OpenAI, Gemini and Anthropic style interfaces and use with tools such as Claude Code and Codex, and it works as an LLM gateway and load balancer. Credentials can be rotated and scheduled across channels, failures can fall back to other credentials, and logs and usage give visibility into what each key or account is used for.

GPT-Load is written in Go with the Gin framework and licensed under MIT. The current 2.0 line is in release-candidate stages, and the README is available in English, Chinese and Japanese. It suits individuals and teams that pool multiple AI provider keys or accounts and want central control.

Key features

Single endpoint for many AI channels
API key and subscription account pooling
Traffic scheduling and load balancing
Failover on credential or channel errors
Request logs and usage accounting
Go-based lightweight server

Good fit for

→Pooling several AI provider keys
→Routing coding agents through one gateway
→Tracking AI usage per credential

Built with: Go
Tags: ai-gateway; llm-gateway; load-balancer; api-gateway; self-hosted; openai; anthropic; golang; failover

GPT-Load: questions and answers

What is GPT-Load used for?: GPT-Load is a self-hosted AI gateway that puts many model channels, API keys and subscription accounts behind a single endpoint with scheduling, failover, request logs and usage accounting. It is a good fit for pooling several AI provider keys, routing coding agents through one gateway and tracking AI usage per credential.
Is GPT-Load open source?: Yes. GPT-Load is open source under the MIT licence. Its source code is on GitHub at tbphp/gpt-load and is written mainly in Go.
Is GPT-Load free?: Yes. GPT-Load is open source, so the software itself is free to use.
Can I self-host GPT-Load?: Yes. GPT-Load can be self-hosted on your own server or infrastructure.
What are some alternatives to GPT-Load?: Similar open-source tools in the AI Infrastructure category include CoAI, New API and LLM Gateway. SaaS products in the same category include AllToken, Costbase and AI Stats.
Is GPT-Load actively maintained?: Yes. The most recent commit to GPT-Load was on 2 October 2026, and the latest release is v2.0.0-rc.41, published on 2 October 2026. The project has 7k stars on GitHub.