Connect your coding workflow.

Claude Code is the recommended connection. It uses the gateway's Anthropic-compatible endpoint and supports streaming, tool use, and multi-turn coding tasks.

export ANTHROPIC_BASE_URL="https://semeru.gaionix.com/api/gateway/anthropic"
export ANTHROPIC_AUTH_TOKEN="gnx_your_api_key"

cd /path/to/your-project
claude --model sonnet

{
  "env": {
    "ANTHROPIC_BASE_URL": "https://semeru.gaionix.com/api/gateway/anthropic",
    "ANTHROPIC_AUTH_TOKEN": "gnx_your_api_key"
  }
}

{
  "enabledPlugins": {
    "figma@claude-plugins-official": true,
    "frontend-design@claude-plugins-official": true
  },
  "model": "sonnet",
  "env": {
    "ANTHROPIC_BASE_URL": "https://semeru.gaionix.com/api/gateway/anthropic",
    "ANTHROPIC_AUTH_TOKEN": "gnx_your_api_key"
  }
}

Choose the endpoint format supported by your application. Both accept your dedicated gnx_... API key.

POST https://semeru.gaionix.com/api/gateway/anthropic/v1/messages
x-api-key: gnx_your_api_key

POST https://semeru.gaionix.com/api/gateway/openai/v1/chat/completions
Authorization: Bearer gnx_your_api_key

Compare estimated usage costs in Gaionix credits before connecting your application.

Cline supports both formats. We recommend selecting Anthropic, enabling Use custom base URL, then setting the Base URL to https://semeru.gaionix.com/api/gateway/anthropic. Enter your gnx_... API key and choose claude-sonnet-4-6 as the model.

Use Cursor's OpenAI base URL override to direct compatible chat requests through Gaionix.

1. 01

   Open Cursor settings

   Open Settings, then find the Models or API Keys section.

2. 02

   Enable the override

   Enable "Override OpenAI Base URL".

3. 03

   Enter Gaionix credentials

   Set the Base URL to https://semeru.gaionix.com/api/gateway/openai/v1 and enter your gnx_ API key.

4. 04

   Select a public model

   Choose a model listed in the OpenAI-compatible table, then save.

Choose one of the models shown in the OpenAI-compatible table above.

OpenClaw runs autonomous AI agents continuously, commonly on a VPS or Docker. It follows standard OpenAI environment variables, so all requests can go through Gaionix without changing your agent workflow.

Environment variables

Create or update the .env file used by OpenClaw. Replace YOUR_PUBLIC_MODEL_ID with the exact model ID from the OpenAI-compatible model table.

# Gaionix OpenAI-compatible gateway
OPENAI_API_BASE=https://semeru.gaionix.com/api/gateway/openai/v1
OPENAI_API_KEY=gnx_your_api_key
OPENAI_MODEL=YOUR_PUBLIC_MODEL_ID

Docker Compose

For a Docker deployment, add the same values under the OpenClaw service environment, then recreate or restart the container.

services:
  openclaw:
    environment:
      OPENAI_API_BASE: "https://semeru.gaionix.com/api/gateway/openai/v1"
      OPENAI_API_KEY: "gnx_your_api_key"
      OPENAI_MODEL: "YOUR_PUBLIC_MODEL_ID"

docker compose up -d
docker compose logs -f openclaw

Choose one of the models shown in the OpenAI-compatible table above.

Both apps connect through Gaionix's OpenAI-compatible endpoint. Use a dedicated API key for the app or workspace so its usage stays easy to review.

# .env
GAIONIX_API_KEY=gnx_your_api_key

# librechat.yaml
endpoints:
  custom:
    - name: "Gaionix"
      apiKey: "${GAIONIX_API_KEY}"
      baseURL: "https://semeru.gaionix.com/api/gateway/openai/v1"
      models:
        default: []
        fetch: true
      titleConvo: true

Credit usage

Credit usage depends on the selected tier and the size of each request and response. Your dashboard shows your remaining credits and request history.