一键导入
deepseek-openclaw-integration
Generate DeepSeek AI model configurations and onboarding commands for OpenClaw agent framework with local-first secret handling
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
菜单
Generate DeepSeek AI model configurations and onboarding commands for OpenClaw agent framework with local-first secret handling
用 Codex 或 Claude 帮你安装 复制这段 Prompt,粘贴到 Codex、Claude 或其他助手里,让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Browser-based interface for viewing and filtering OpenClaw session tool call history with zero dependencies for local network deployment.
AI-powered quantitative research and backtesting platform with end-to-end workflow from research to strategy publication
Give your AI assistant a phone — OpenClaw plugin for real phone calls via Twilio + OpenAI Realtime API with in-call tools, transcripts, and call screening
Run multi-model consensus panels (Lite or Heavy) with your own agent backends—no hosted middleware, your models, your rules.
Build a multi-role JARVIS-style voice assistant with local ASR/TTS, OpenClaw LLM gateway, voice wake words, HUD effects, and speaker verification
Use 37 battle-tested marketing skills covering CRO, copywriting, SEO, paid ads, email, growth, and strategy with real data connectors for Google Ads, Search Console, Meta Ads, and X/Twitter
| name | deepseek-openclaw-integration |
| description | Generate DeepSeek AI model configurations and onboarding commands for OpenClaw agent framework with local-first secret handling |
| triggers | ["how do I integrate DeepSeek with OpenClaw","set up DeepSeek models in OpenClaw","generate OpenClaw configuration for DeepSeek","connect DeepSeek API to OpenClaw","configure DeepSeek v4 flash in OpenClaw","create OpenClaw agent with DeepSeek reasoner","DeepSeek OpenClaw onboarding command","switch OpenClaw to DeepSeek models"] |
Skill by ara.so — Hermes Skills collection.
DeepSeek OpenClaw is a local-first React application that helps developers integrate DeepSeek AI models into the OpenClaw agent framework. It generates configuration snippets, onboarding commands, and provides model selection tooling without exposing API keys to backend services.
git clone https://github.com/Chiptreevaluate/deepseek-openclaw-863.git
cd deepseek-openclaw-863
npm install
npm start
npm run dev
npm run build
npm run preview
| Model ID | Purpose |
|---|---|
deepseek/deepseek-v4-flash | Fast execution for everyday automation and quick responses |
deepseek/deepseek-v4-pro | Complex coding tasks, multi-step planning |
deepseek/deepseek-chat | General conversational interactions without reasoning overhead |
deepseek/deepseek-reasoner | Deep reasoning and step-by-step problem solving |
OpenClaw agents use environment variables for API authentication and JSON configuration for model selection. DeepSeek OpenClaw generates both:
The tool generates OpenClaw-compatible configuration:
{
"env": {
"DEEPSEEK_API_KEY": "${DEEPSEEK_API_KEY}"
},
"agents": {
"defaults": {
"model": {
"primary": "deepseek/deepseek-v4-flash"
}
}
}
}
Before running generated commands, set your DeepSeek API key:
# Linux/macOS
export DEEPSEEK_API_KEY=your_key_here
# Windows PowerShell
$env:DEEPSEEK_API_KEY="your_key_here"
# Windows CMD
set DEEPSEEK_API_KEY=your_key_here
The tool produces commands following this pattern:
openclaw onboard \
--provider deepseek \
--model deepseek/deepseek-v4-flash \
--api-key ${DEEPSEEK_API_KEY}
// src/models.ts
export interface DeepSeekModel {
id: string;
name: string;
description: string;
useCase: string;
}
export const DEEPSEEK_MODELS: DeepSeekModel[] = [
{
id: 'deepseek/deepseek-v4-flash',
name: 'DeepSeek v4 Flash',
description: 'Fast model for everyday automation',
useCase: 'Quick responses, simple tasks'
},
{
id: 'deepseek/deepseek-v4-pro',
name: 'DeepSeek v4 Pro',
description: 'Advanced model for complex reasoning',
useCase: 'Multi-step planning, complex coding'
},
{
id: 'deepseek/deepseek-chat',
name: 'DeepSeek Chat',
description: 'General conversational model',
useCase: 'Chat interfaces, simple Q&A'
},
{
id: 'deepseek/deepseek-reasoner',
name: 'DeepSeek Reasoner',
description: 'Step-by-step reasoning model',
useCase: 'Logic puzzles, debugging, analysis'
}
];
// src/openclaw.ts
export interface OpenClawConfig {
env: Record<string, string>;
agents: {
defaults: {
model: {
primary: string;
};
};
};
}
export function generateOpenClawConfig(
modelId: string,
apiKey?: string
): OpenClawConfig {
return {
env: {
DEEPSEEK_API_KEY: apiKey || '${DEEPSEEK_API_KEY}'
},
agents: {
defaults: {
model: {
primary: modelId
}
}
}
};
}
export function generateOnboardingCommand(
modelId: string,
useEnvVar: boolean = true
): string {
const keyRef = useEnvVar ? '${DEEPSEEK_API_KEY}' : '<your-api-key>';
return `openclaw onboard --provider deepseek --model ${modelId} --api-key ${keyRef}`;
}
// src/App.tsx
import React, { useState } from 'react';
import { DEEPSEEK_MODELS } from './models';
import { generateOpenClawConfig, generateOnboardingCommand } from './openclaw';
export default function App() {
const [selectedModel, setSelectedModel] = useState(DEEPSEEK_MODELS[0].id);
const [apiKey, setApiKey] = useState('');
const config = generateOpenClawConfig(selectedModel, apiKey);
const command = generateOnboardingCommand(selectedModel, !apiKey);
const redactedConfig = {
...config,
env: {
DEEPSEEK_API_KEY: apiKey ? 'sk-***redacted***' : '${DEEPSEEK_API_KEY}'
}
};
return (
<div>
<h1>DeepSeek OpenClaw Configuration</h1>
<select
value={selectedModel}
onChange={(e) => setSelectedModel(e.target.value)}
>
{DEEPSEEK_MODELS.map(model => (
<option key={model.id} value={model.id}>
{model.name}
</option>
))}
</select>
<input
type="password"
placeholder="DeepSeek API Key (optional)"
value={apiKey}
onChange={(e) => setApiKey(e.target.value)}
/>
<h3>Onboarding Command</h3>
<pre>{command}</pre>
<h3>Config Snippet</h3>
<pre>{JSON.stringify(redactedConfig, null, 2)}</pre>
</div>
);
}
Always use environment variables for secrets:
// ✅ Good: Reference environment variable
const config = {
env: {
DEEPSEEK_API_KEY: '${DEEPSEEK_API_KEY}'
}
};
// ❌ Bad: Hardcoded secret
const config = {
env: {
DEEPSEEK_API_KEY: 'sk-actual-key-value'
}
};
// Choose model based on task complexity
function selectModel(taskType: string): string {
switch (taskType) {
case 'quick-automation':
return 'deepseek/deepseek-v4-flash';
case 'complex-coding':
return 'deepseek/deepseek-v4-pro';
case 'reasoning':
return 'deepseek/deepseek-reasoner';
default:
return 'deepseek/deepseek-chat';
}
}
{
"agents": {
"defaults": {
"model": {
"primary": "deepseek/deepseek-v4-flash"
}
},
"coding-assistant": {
"model": {
"primary": "deepseek/deepseek-v4-pro"
}
},
"reasoning-agent": {
"model": {
"primary": "deepseek/deepseek-reasoner"
}
}
}
}
Use the UI or programmatically generate config:
import { generateOpenClawConfig } from './openclaw';
const config = generateOpenClawConfig('deepseek/deepseek-v4-flash');
console.log(JSON.stringify(config, null, 2));
export DEEPSEEK_API_KEY=your_actual_key
openclaw onboard \
--provider deepseek \
--model deepseek/deepseek-v4-flash \
--api-key ${DEEPSEEK_API_KEY}
openclaw test --model deepseek/deepseek-v4-flash
Problem: OpenClaw cannot authenticate with DeepSeek.
Solution: Verify environment variable is set correctly:
# Check if variable exists
echo $DEEPSEEK_API_KEY
# Re-export if needed
export DEEPSEEK_API_KEY=your_key_here
Problem: Selected model ID is not recognized by OpenClaw.
Solution: Verify model ID format matches OpenClaw provider conventions:
// Correct format
const modelId = 'deepseek/deepseek-v4-flash';
// Check against available models
import { DEEPSEEK_MODELS } from './models';
const isValid = DEEPSEEK_MODELS.some(m => m.id === modelId);
Problem: Generated config doesn't affect OpenClaw behavior.
Solution: Ensure config is saved to OpenClaw's expected location:
# Default OpenClaw config location
~/.openclaw/config.json
# Or use --config flag
openclaw run --config ./my-deepseek-config.json
Problem: Antivirus blocks npm install or execution.
Solution: The project is open source with MIT license. If security software blocks:
Problem: TypeScript or Vite build fails.
Solution: Ensure dependencies are installed and versions match:
# Clean install
rm -rf node_modules package-lock.json
npm install
# Check Node version (requires Node 16+)
node --version
// ✅ Use environment variables
const apiKey = process.env.DEEPSEEK_API_KEY;
// ✅ Redact in logs and UI
const redactedKey = apiKey ? 'sk-***redacted***' : '<not-set>';
// ❌ Never log actual keys
console.log('Using key:', apiKey); // Don't do this
# Add to .gitignore
echo "*.key" >> .gitignore
echo ".env.local" >> .gitignore
echo "openclaw.config.json" >> .gitignore
Before committing generated configs:
# Search for potential secrets
git diff | grep -E "sk-[a-zA-Z0-9]{32,}"
deepseek-openclaw-863/
├── src/
│ ├── App.tsx # Main React component
│ ├── models.ts # DeepSeek model definitions
│ ├── openclaw.ts # Config generation logic
│ ├── types.ts # TypeScript interfaces
│ └── styles.css # UI styles
├── .env.example # Environment variable template
├── package.json # Dependencies and scripts
├── tsconfig.json # TypeScript configuration
└── vite.config.ts # Vite build configuration