Complete Tutorial for Connecting Moltbot to an API Proxy: 5 Steps to Configure OpenAI Compatible Interfaces and Save 60% Cost

By default, Moltbot uses the official Anthropic API, but the official service comes with certain access restrictions and higher costs. In this article, we'll walk you through how to configure Moltbot to connect to an API relay, allowing you to use third-party API services via the openai-compatible mode.

Core Value: After reading this, you'll know how to set up Moltbot with an API relay for a more affordable and stable AI assistant experience.

Key Points for Moltbot API Relay Configuration

Before diving into the setup, let's look at Moltbot's API configuration mechanism and why using a relay is beneficial.

Comparison of Moltbot API Configuration Methods

Moltbot API Proxy Configuration: Getting Started

Step 1: Confirm Moltbot is installed

First, make sure you've got Moltbot installed on your machine:

# Check your Moltbot version
moltbot --version

# If it's not installed, run the installation
npm install -g moltbot@latest

System Requirements: Node.js >= 22

Step 2: Get your API Proxy Key

You'll need to grab an API Key from an API proxy provider.

🚀 Quick Start: We recommend using APIYI (apiyi.com) to get your API keys. You'll get free credits just for signing up. It supports all the heavy hitters like Claude Sonnet 4, Claude Opus 4.5, and GPT-4o, with prices typically 40-60% lower than official rates.

Once you have your key, jot down these details:

Method 1: Configuring via config.json

This is the recommended way to go—configure it once, and you're set for good.

Finding the Moltbot Config File

The config file usually lives in one of these spots:

You can also find it using the command line:

# See your current config
moltbot config list

# Get the exact path to your config file
moltbot config path

Editing config.json for the API Proxy

Open up your config file and add or update the models.providers section like this:

{
  "models": {
    "providers": {
      "apiyi": {
        "baseUrl": "https://api.apiyi.com/v1",
        "apiKey": "sk-your-api-key-here",
        "api": "openai-completions",
        "authHeader": true,
        "models": [
          {
            "id": "claude-sonnet-4-20250514",
            "name": "Claude Sonnet 4",
            "contextWindow": 200000,
            "maxTokens": 64000
          },
          {
            "id": "claude-opus-4-5-20251101",
            "name": "Claude Opus 4.5",
            "contextWindow": 200000,
            "maxTokens": 32000
          },
          {
            "id": "gpt-4o",
            "name": "GPT-4o",
            "contextWindow": 128000,
            "maxTokens": 16384
          }
        ]
      }
    }
  },
  "agent": {
    "model": {
      "primary": "apiyi/claude-sonnet-4-20250514",
      "fallbacks": ["apiyi/claude-opus-4-5-20251101", "apiyi/gpt-4o"]
    }
  }
}

Configuration Parameters Explained

Moltbot API Proxy Configuration Method 2: YAML Format

If you prefer the YAML format, Moltbot also supports config.yaml:

# ~/.moltbot/config.yaml
models:
  providers:
    apiyi:
      baseUrl: "https://api.apiyi.com/v1"
      apiKey: "sk-your-api-key-here"
      api: openai-completions
      authHeader: true
      models:
        - id: claude-sonnet-4-20250514
          name: Claude Sonnet 4
          contextWindow: 200000
          maxTokens: 64000
        - id: claude-opus-4-5-20251101
          name: Claude Opus 4.5
          contextWindow: 200000
          maxTokens: 32000

agent:
  model:
    primary: apiyi/claude-sonnet-4-20250514
    fallbacks:
      - apiyi/claude-opus-4-5-20251101
      - apiyi/gpt-4o

Simplified YAML Configuration

If you only need to use a single model, you can use a more streamlined configuration:

# ~/.moltbot/config.yaml - 极简版
llm:
  provider: openai-compatible
  model: claude-sonnet-4-20250514
  apiKey: sk-your-api-key-here
  baseUrl: https://api.apiyi.com/v1

🎯 Technical Tip: We recommend using the full configuration. This way, you can set up multiple models and fallback strategies to improve service stability. Through APIYI (apiyi.com), you can simultaneously access various models like Claude, GPT, and Gemini.

Moltbot API Proxy Configuration Method 3: Environment Variables

This method is ideal for temporary testing or CI/CD environments:

# 设置环境变量
export MOLTBOT_LLM_PROVIDER="openai-compatible"
export MOLTBOT_LLM_MODEL="claude-sonnet-4-20250514"
export MOLTBOT_LLM_API_KEY="sk-your-api-key-here"
export MOLTBOT_LLM_BASE_URL="https://api.apiyi.com/v1"

# 启动 Moltbot
moltbot gateway start

Alternatively, you can configure them in a .env file:

# ~/.moltbot/.env
MOLTBOT_LLM_PROVIDER=openai-compatible
MOLTBOT_LLM_MODEL=claude-sonnet-4-20250514
MOLTBOT_LLM_API_KEY=sk-your-api-key-here
MOLTBOT_LLM_BASE_URL=https://api.apiyi.com/v1

Moltbot API Proxy Configuration Method 4: CLI Configuration

You can quickly configure everything using the moltbot config set command:

# Configure API proxy
moltbot config set models.providers.apiyi.baseUrl "https://api.apiyi.com/v1"
moltbot config set models.providers.apiyi.apiKey "sk-your-api-key-here"
moltbot config set models.providers.apiyi.api "openai-completions"

# Set default model
moltbot config set agent.model.primary "apiyi/claude-sonnet-4-20250514"

# Verify configuration
moltbot config list

Batch Configuration Script

If you'd rather do it all at once, create a configuration script to handle the setup in one click:

#!/bin/bash
# setup-apiyi.sh - Moltbot APIYI Proxy Configuration Script

API_KEY="sk-your-api-key-here"
BASE_URL="https://api.apiyi.com/v1"

echo "Configuring Moltbot API Proxy..."

# Set up provider
moltbot config set models.providers.apiyi.baseUrl "$BASE_URL"
moltbot config set models.providers.apiyi.apiKey "$API_KEY"
moltbot config set models.providers.apiyi.api "openai-completions"
moltbot config set models.providers.apiyi.authHeader true

# Set default model
moltbot config set agent.model.primary "apiyi/claude-sonnet-4-20250514"

# Restart gateway
moltbot gateway restart

echo "Configuration complete! Use 'moltbot doctor' to verify your settings."

Verifying Your Moltbot API Proxy Configuration

After you've finished the setup, you'll want to make sure everything is running correctly.

Check with moltbot doctor

# Run diagnostics
moltbot doctor

# Automatically fix issues
moltbot doctor --fix

Here's what a healthy output looks like:

✓ Gateway running on port 18789
✓ Model provider 'apiyi' configured
✓ API key validated
✓ Model 'claude-sonnet-4-20250514' available
✓ Connection test passed

Send a Test Message

# Test API connection
moltbot test-llm --provider apiyi --model claude-sonnet-4-20250514

# Send a test chat message
moltbot chat "Hello, please introduce yourself."

Check API Call Logs

If something isn't working right, you can dive into the logs:

# View real-time logs
moltbot logs --follow

# View the last 10 API calls
moltbot logs --filter api-call --last 10

Models Supported by the Moltbot APIYI Proxy

Through the APIYI proxy, you can use the following models in Moltbot:

Claude Series Models

OpenAI Series Models

Google Series Models

💡 Recommendation: For daily Moltbot use, we recommend Claude Sonnet 4. It strikes a great balance between performance and cost. You can get the full model list and real-time pricing via APIYI at apiyi.com.

Model Ecosystem Supported by Moltbot via APIYI One API Key · Multi-provider Models · Unified Interface

APIYI apiyi.com Unified Proxy

Anthropic Claude • Opus 4.5 Strongest Reasoning • Sonnet 4 Recommended • Haiku 3.5 Fast Response Context: 200K tokens

OpenAI GPT • GPT-4o Multimodal • GPT-4o Mini Lightweight & Fast • o1 Series Deep Reasoning Context: 128K tokens

Google Gemini • Gemini 2.0 Pro Pro • Gemini 2.0 Flash Recommended • Gemini 1.5 Pro Stable Context: 1M tokens

More Models • DeepSeek Open Source Powerhouse • Llama 3 Meta Open Source • Mistral Light of Europe Continuously updating…

🎯 Recommended: Claude Sonnet 4 – Best performance/cost balance

Moltbot API Proxy Advanced Configuration

Configuring Fallback Strategies

Automatically switch to a backup model when the primary model is unavailable:

{
  "agent": {
    "model": {
      "primary": "apiyi/claude-sonnet-4-20250514",
      "fallbacks": [
        "apiyi/claude-3-5-sonnet-20241022",
        "apiyi/gpt-4o"
      ],
      "fallbackStrategy": "sequential"
    }
  }
}

Configuring Model Routing Rules

Automatically select the best model based on the task type:

{
  "agent": {
    "modelRouting": {
      "coding": "apiyi/claude-sonnet-4-20250514",
      "reasoning": "apiyi/claude-opus-4-5-20251101",
      "quick": "apiyi/claude-3-5-haiku-20241022",
      "default": "apiyi/claude-sonnet-4-20250514"
    }
  }
}

Configuring Cost Limits

Prevent unexpected high spending:

{
  "agent": {
    "limits": {
      "maxTokensPerRequest": 32000,
      "maxRequestsPerHour": 100,
      "maxCostPerDay": 10.00
    }
  }
}

Moltbot API Proxy FAQs

# Verify API Key
curl -H "Authorization: Bearer sk-your-key" https://api.apiyi.com/v1/models

moltbot config set agent.model.primary "apiyi/claude-opus-4-5-20251101"
moltbot gateway restart

moltbot chat --model apiyi/gpt-4o "Hello"

{
  "agent": {
    "model": {
      "primary": "apiyi/claude-sonnet-4-20250514",
      "fallbacks": ["openrouter/anthropic/claude-3.5-sonnet"]
    }
  }
}

# View local statistics
moltbot stats --period today

# View detailed usage
moltbot stats --detailed --period week

Moltbot API Proxy Troubleshooting

Common Errors & Solutions

Debug Mode

Enable detailed logging to help pinpoint issues:

# Enable debug mode
export MOLTBOT_DEBUG=true
moltbot gateway start

# Or use command line arguments
moltbot gateway start --debug --log-level verbose

Configuration Validation Commands

# Validate config syntax
moltbot config validate

# Test API connection
moltbot test-connection --provider apiyi

# Full health check
moltbot doctor --verbose

Moltbot API Proxy vs. Official API

🎯 Recommendation: For users in mainland China or those who need multi-model support, we recommend connecting to Moltbot via APIYI (apiyi.com). You'll get better pricing, more stable service, and dedicated technical support in Chinese.

Summary: Recapping Key Moltbot API Proxy Configuration Points

By following this tutorial, you've learned how to fully configure Moltbot to work with an API proxy:

1. Preparation: Install Moltbot and get your API Key from APIYI (apiyi.com).
2. Configuration Files: Edit config.json or config.yaml to add your provider settings.
3. Key Parameters: baseUrl, apiKey, api type, and the list of models.
4. Verification & Testing: Use moltbot doctor and test commands to confirm everything's working correctly.
5. Advanced Features: Fallback strategies, model routing, and cost limits.

We recommend getting your API keys through APIYI (apiyi.com) to enjoy lower prices and more stable service, helping you get the most out of your Moltbot assistant.

References

1. Official Moltbot Config Docs: Gateway configuration reference

   • Link: docs.molt.bot/gateway/configuration

2. Moltbot Configuration Examples: Templates for various scenarios

   • Link: docs.molt.bot/gateway/configuration-examples

3. Moltbot Model Providers: Detailed provider configuration guide

   • Link: docs.molt.bot/concepts/model-providers

4. Moltbot GitHub: Source code and issues

   • Link: github.com/moltbot/moltbot

5. Moltbot Getting Started Guide: Quick start tutorial

   • Link: docs.molt.bot/start/getting-started

📝 Author: APIYI Team
🔗 Technical Support: If you need API keys or technical help, feel free to visit APIYI (apiyi.com).