OpenClaw + Step 3.5 Flash Setup Guide

Step 3.5 Flash를 사용하여 MacOS에 OpenClaw를 설치, 구성 및 배포하세요.

필수 조건

• 운영체제: macOS (Apple Silicon/Intel)
• 실행 환경: Node.js (npm을 사용하는 경우 선택 사항)

설치

OpenClaw를 설치하려면 다음 방법 중 하나를 선택하십시오.

옵션 1: 공식 스크립트 (권장)

curl -fsSL https://openclaw.ai/install.sh | bash

옵션 2: NPM

npm i -g openclaw
# 권한 오류가 발생하면 sudo를 사용하세요.

빠른 시작: 온보딩

설정 마법사를 사용하여 상담원을 초기화하세요.

1. 온보드 명령어를 실행하세요:

   openclaw onboard
   

2. 마법사를 따라가세요:

   • 환영합니다: '예'를 선택하세요.
   • 모드: '빠른 시작'을 선택하세요.
   • 모델 구성: '지금은 건너뛰기'를 선택하세요(3.5단계 플래시 구성은 웹 UI에서 진행합니다).
   • 통합: '지금은 건너뛰기'를 선택하세요(필요한 경우 나중에 Telegram을 구성하세요).
   • 완료: 기본 설정을 유지합니다('Enter' 키를 누르세요).
   📷 온보딩 스크린샷 보기

   Step	Screenshot
   Welcome
   Quick Start
   Skip Model
   Complete

3. Launch WebUI:
   If prompted, select WebUI to open the dashboard. If missed, run:

   openclaw dashboard
   

Configuring Step 3.5 Flash

You can configure the model via the WebUI (Recommended) or by editing the JSON config.

Method 1: WebUI Configuration (Recommended)

1. Open Settings: Navigate to Config -> Models -> Providers.
   

2. Add Provider:

   • Click Add Entry.
   • Name: stepfun (or custom-1).
   • API Type: openai-completions.
   • Base URL:
     • StepFun (International): https://api.stepfun.ai/v1
     • StepFun (China): https://api.stepfun.com/v1
     • OpenRouter: https://openrouter.ai/api/v1
   • API Key: Enter your StepFun API Key.
     

3. Add Model:

   • Scroll to the Models section and click Add.
   • ID: step-3.5-flash (or stepfun/step-3.5-flash for OpenRouter).
   • Name: Step 3.5 Flash.
   • Context Window: 256000 (This is the maxium context window size for Step 3.5 Flash).
     

4. Set as Primary:

   • Search for primary in the settings.
   • Set value to "stepfun/step-3.5-flash".
   • Save & Reload: Click the Save icon 💾, then the Reload icon 🔄.
     

Method 2: Manual JSON Configuration

Edit ~/.openclaw/openclaw.json directly.

📄 Click to view `openclaw.json` template

{
  "models": {
    "providers": {
      "stepfun": {
        "baseUrl": "https://api.stepfun.ai/v1", // OR https://api.stepfun.com/v1 (China) OR https://openrouter.ai/api/v1
        "apiKey": "YOUR_SK_KEY_HERE",
        "auth": "api-key",
        "api": "openai-completions",
        "models": [
          {
            "id": "step-3.5-flash", // OR stepfun/step-3.5-flash
            "name": "Step 3.5 Flash",
            "contextWindow": 256000,
            "maxTokens": 8192
          }
        ]
      }
    }
  },
  "agents": {
    "defaults": {
      "model": {
        "primary": "stepfun/step-3.5-flash"
      }
    }
  }
}

Integrations (Optional)

Telegram Bot

1. Create Bot: Chat with BotFather, run /newbot to get a Token.
2. Get User ID: Chat with userinfobot to get your ID.
3. Configure: In openclaw onboard (or GUI), input the Token and ID.

Command Reference

Command	Description
openclaw onboard	Re-run setup wizard.
openclaw dashboard	Open the WebUI.
openclaw gateway	Start background service (Telegram/WhatsApp).
openclaw gui	Open native desktop app.

Troubleshooting

WebUI Won't Open
Run openclaw onboard again to ensure the gateway is initialized correctly.

Model Not Responding

1. Check Error Logs in WebUI.
2. Verify apiKey and baseUrl are correct.
3. Use the connection test script below.

🐍 Python Connection Test Script

import os
from openai import OpenAI

# Configuration
MODEL = "step-3.5-flash" 
API_KEY = "YOUR_API_KEY"      
BASE_URL = "https://api.stepfun.ai/v1" # OR https://api.stepfun.com/v1 (China) OR https://openrouter.ai/api/v1

client = OpenAI(api_key=API_KEY, base_url=BASE_URL)

try:
    print("Testing connection...")
    resp = client.chat.completions.create(
        model=MODEL,
        messages=[{"role": "user", "content": "Hello, are you working?"}],
        stream=False
    )
    print(f"Response: {resp.choices[0].message.content}")
except Exception as e:
    print(f"Error: {e}")