OpenViSTAI · Agents · Blockchain · Crypto
Submit a project
Menu
AI Infrastructure & Serving

Gpt Load

Gpt Load is a self-hosted AI gateway that puts multiple provider channels, API keys and subscription accounts behind one entry point. It schedules traffic across configured credentials, handles failures, and provides request logs and usage accounting.

Explore the practical guide ↓

Independent OpenViST project profile, based on linked sources.

Gpt Load — OpenViST project profile in AI Infrastructure & Serving.
OpenViST project illustration. See the linked sources for product details.
Understand the project

Practical overview

When it fits

Use Gpt Load when you want to operate a self-hosted gateway between client applications and multiple AI provider channels, while retaining native OpenAI, Anthropic or Gemini interfaces for clients.

Who it is for

  • Developers who need a single gateway endpoint for applications using multiple AI service credentials.
  • Teams operating their own AI routing infrastructure and managing provider, model and credential settings.

Documented uses

  • Give client applications one base URL and AccessKey for configured AI services.
  • Distribute requests across credentials using scheduling and weights.
  • Apply retries, cooldowns, blacklisting and session affinity to manage credential failures and load.
  • Review gateway health, routes, request logs, usage and cost estimates.
Before you choose

Requirements & limits

The project is open source under the MIT license. Its documented deployment model is self-hosted.

  • The gateway is deployed and operated by the project user rather than used as a hosted model service.
  • Provider credentials, models and routing policies are configured through its management interface.
  • Documented storage options are SQLite, MySQL and PostgreSQL; the project also describes local credential encryption.
  • Clients can keep native OpenAI, Anthropic or Gemini interfaces while connecting through the gateway.

Check current requirements in the official documentation. Source-based guidance does not mean OpenViST has run or independently tested this project.

Source-based practical guide

Start GPT-Load with Docker Compose and complete initial gateway configuration

Start the service, sign in to its console, and configure a channel, group, and AccessKey.

Before you start

  • Docker and Docker Compose

Steps

  1. Clone the repository, enter its directory, copy .env.example to .env, and start the service with the command below.

    Documented command
    docker compose up -d

  2. Check service health at http://127.0.0.1:3001/health, then retrieve the generated management key with the command below.

    Documented command
    docker compose exec gpt-load sh -c 'cat /app/data/auth.key'

  3. Open http://127.0.0.1:3001 and sign in to the console using the management key.

  4. Add a channel and credentials, complete OAuth or import credentials if using a subscription channel, create a group with its models and runtime policy, then create an AccessKey with allowed groups and client protocols.

Expected outcome

GPT-Load is running and the console has a configured upstream group and an AccessKey that can be given to an application.

Guide limitations

  • The service listens on the loopback address by default and is not exposed to the internet.
  • For subscription channels, OAuth clients use fixed callback ports; only one default Compose instance can run on a host at a time.
  • With SSH or a remote browser, the browser's localhost may not reach GPT-Load; paste the full callback URL into the authorization dialog.
  • Version 2.0 cannot open, import, or migrate 1.x data in place.
Documentation 1 ↗Documentation 2 ↗
Evidence and transparency

Sources & review

Official repository ↗Guide documentation ↗
Source record details
Source availability
Active
Sources checked
7 October 2026
Original website title
tbphp/gpt-load · Official homepage title · retained in its source language
How we review sourcesSuggest a correction