# tbphp/gpt-load

Multi-channel AI proxy with intelligent key rotation. 智能密钥轮询的多渠道 AI 代理。

Repository: https://github.com/tbphp/gpt-load
Canonical: https://ross.abutalabs.com/products/gpt-load
Homepage: https://www.gpt-load.com
Language: Go
License: MIT
License Family: permissive
Topics: go, golang, openai, gemini, gin, anthropic, claude
Last push: 2026-08-26T15:01:15+00:00

## Health v2 (maintenance only)
Score: 83/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 99, release rhythm 91, longevity 32
- inputs: {"age_days": 453, "days_push": 7, "days_rel": 62, "gap_med": 6.0, "n_releases_24m": 33}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 6290, forks 674 (observed 2026-08-28T04:09:41.359122+00:00)

## What it is
GPT-Load is a high-performance, enterprise-grade AI API transparent proxy service built in Go that aggregates multiple AI providers (OpenAI, Google Gemini, Anthropic Claude) behind a unified API. It provides intelligent API key pool management with automatic rotation, load balancing, and a web-based admin dashboard with monitoring, deployable standalone via Docker or as a master/slave cluster.

## Use cases
- rotate multiple OpenAI API keys automatically to avoid rate limits
- proxy OpenAI, Gemini, and Claude behind one unified endpoint
- load balance LLM API requests across multiple providers
- self-host an AI gateway with key pool management and monitoring
- deploy a high-concurrency AI proxy cluster with master/slave architecture
- track LLM API usage and request logs in a dashboard

## When to choose
- you need to manage many API keys across multiple AI providers with automatic failover
- you want a transparent proxy that preserves native OpenAI/Gemini/Claude API formats
- you need production-grade high-concurrency proxying with clustering and monitoring
- you want a self-hosted alternative with a web admin UI and Docker deployment

## When to avoid
- you only use a single AI provider with one API key and no rate-limit issues
- you need request transformation, routing rules, or advanced gateway features beyond transparent proxying
- you want a fully managed cloud gateway rather than self-hosting with MySQL/Redis infrastructure

## Facets
- artifact type: service
- maturity: active
- function: proxy, api-gateway, http-server, monitoring, caching, load-testing
- domain: artificial-intelligence, large-language-models, apis, self-hosted, backend
- platform: self-hosted, windows, go
- tags: ai-gateway, key-rotation, openai-proxy, gemini, anthropic, load-balancing, transparent-proxy, multi-channel, devops, docker, linux, macos, web-server

## Member repositories
- tbphp/gpt-load (main) score 83

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:09:41.359122+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T17:46:18.311854+00:00, confidence not recorded.
  - readme: https://github.com/tbphp/gpt-load (fetched 2026-08-28T04:09:41.359122+00:00, sha d8ebf71fd464)
  - homepage: https://www.gpt-load.com (fetched 2026-08-29T08:42:57.690940+00:00, sha fe5968ec6c05)
  - site_page: https://www.gpt-load.com/docs (fetched 2026-08-29T08:42:57.694778+00:00, sha 2f6c2d63a436)
  - site_page: https://www.gpt-load.com/docs/deployment (fetched 2026-08-29T08:42:57.696465+00:00, sha 1246bb32714b)
  - site_page: https://www.gpt-load.com/docs/deployment/standalone (fetched 2026-08-29T08:42:57.698055+00:00, sha add44427eab6)
  - site_page: https://www.gpt-load.com/docs/sponsor (fetched 2026-08-29T08:42:57.699745+00:00, sha db3aee858509)
  - site_page: https://www.gpt-load.com/docs/deployment/cluster (fetched 2026-08-29T08:42:57.701235+00:00, sha db6a864e7ed7)
- Data as of 2026-08-30T08:39:29.467469+00:00.