Panduan Provider

Guide to registering as an LLM API provider on ARZASTORE AI Marketplace

Earn Revenue from APIs

Create a Provider

1. 01Log in → Console → LLM → My Providers
2. 02Click Create New Provider
3. 03Fill in provider details:Provider Name: Display name on the Marketplace (e.g. "My OpenAI API")Provider Type: Choose the API type (OpenAI, Claude, Gemini, Ollama, OpenAI-Compatible)Base URL: Original API endpoint (e.g. https://api.openai.com/v1)Description: Short description about your provider (shown on Marketplace)

Provider Name: Display name on the Marketplace (e.g. "My OpenAI API")

Provider Type: Choose the API type (OpenAI, Claude, Gemini, Ollama, OpenAI-Compatible)

Base URL: Original API endpoint (e.g. https://api.openai.com/v1)

Description: Short description about your provider (shown on Marketplace)

Supported Provider Types

Official API from OpenAI (GPT-4o, GPT-4.1, o3, DALL-E, etc.)

Official API from Anthropic (Claude Haiku, Sonnet, Opus)

API from Google (Gemini 2.5 Pro, Flash, etc.)

Self-hosted models (Llama, Mistral, CodeLlama, etc.)

Any OpenAI-compatible API (vLLM, Together, Groq, etc.)

Add API Keys

Add API keys from the original AI service to the provider. You can add multiple keys so the system automatically distributes requests (load balancing).

1. 01Open the provider detail drawer → "API Keys" tab
2. 02Click Add Key, paste the API key from the original provider
3. 03Configure options:Key Name: Description (e.g. "OpenAI Key 1", "Backup Key")Priority: Priority order (higher-priority keys are used first)Available Models: Restrict the key to specific models (optional)

Key Name: Description (e.g. "OpenAI Key 1", "Backup Key")

Priority: Priority order (higher-priority keys are used first)

Available Models: Restrict the key to specific models (optional)

Multiple Keys = Load Balancing

Configure Models

Add and configure the models your provider supports. Each model has its own pricing.

1. 01"Models" tab → Add Model or Auto-fetch list
2. 02Fill in model details:Model ID: Exact model name (e.g. gpt-4o, claude-sonnet-4-20250514)Display Name: Display name on the MarketplaceContext Length: Context window size (e.g. 128000 tokens)Max Output Tokens: Maximum tokens for output
3. 03Configure pricing:

Model ID: Exact model name (e.g. gpt-4o, claude-sonnet-4-20250514)

Display Name: Display name on the Marketplace

Context Length: Context window size (e.g. 128000 tokens)

Max Output Tokens: Maximum tokens for output

Suitable for chat/completion models. Priced per input/output token count.

Input Price: IDR / 1M tokens

Output Price: IDR / 1M tokens

Example GPT-4o pricing:

Input: 3,000 IDR / 1M tokens

Output: 12,000 IDR / 1M tokens

Suitable for models with fixed pricing or self-hosted. Fixed price per call.

Request Price: IDR / request

No token limit

Example Llama (self-hosted) pricing:

Price: 50 IDR / request

Auto-fetch Model List

Validate & Activate

Before the provider appears on the Marketplace, the system will validate API keys and model configuration.

1. 01Click Validate in the provider drawer
2. 02The system will send test requests to the original API via each key → check connectivity
3. 03If validation succeeds → provider will have status Active and appear on the Marketplace

Automatic Health Check

The system automatically checks provider health every 5 minutes. If API keys are unresponsive or return repeated errors, the provider will be markedUnhealthy and temporarily hidden from the Marketplace.

API working well

Some keys failing

All keys failing

Earnings & Withdrawal

Every time a consumer calls your provider's API, you receive corresponding earnings (minus the platform fee).

How Earnings are Calculated

Example:

GPT-4o output price: 12,000 IDR / 1M tokens

Consumer uses 100K output tokens → Total: 1,200 IDR

Platform fee 10% → 120 IDR

→ Your earnings: 1,080 IDR

Track Earnings

• Real-time dashboard showing total earnings
• Statistics by model, by day/week/month
• Per-request details: tokens, cost, time
• Per-key stats: see which key gets the most usage

Withdrawal

• Withdraw earnings from Console → LLM → Earnings
• Earnings are transferred to your main wallet
• From the main wallet, you can withdraw to your bank
• Minimum withdrawal: see rules on the withdrawal page

Best Practices

Add Multiple API Keys

Adding 2–3 keys per provider helps the system round-robin between keys, avoid rate limits, and increase throughput. If one key runs out of quota, another automatically takes over.

Set Competitive Prices

Check other providers' prices on the Marketplace to set competitive rates. Consumers can use the cheapest strategy → lower-priced providers get priority.

Monitor Health Status

Check the Dashboard regularly. If a key becomes Unhealthy, check quota/billing at the original provider and replace the key if needed.

Offer Diverse Models

Add both budget models (GPT-4o-mini) and powerful models (GPT-4o, Claude Opus) to serve different consumer needs.

Troubleshooting

Issue	Cause	Fix
Validation failed	Wrong or expired API key	Check key at original provider, replace with new key
Health check: Unhealthy	Key out of quota, billing expired	Top up at provider, add backup key
Model not showing on Marketplace	Provider not active or model not enabled	Validate provider, check model is enabled
Consumer reports 502 error	Original API timeout or down	Check provider status page, add backup key
Earnings not increasing	No consumers subscribed yet	Set competitive prices, add more models

Summary: 5 Steps to Become a Provider