Real-Time Search in OpenClaw Agents with Scavio Skills

OpenClaw agents can reason, plan, and execute tasks -- but by default they have no window to the live web. Ask one about a product released last month, a trending video, or today's deals on Amazon and it will either hallucinate or admit it does not know. Adding real-time search changes that. This guide walks through installing Scavio's four search skills from ClawHub and wiring them into an OpenClaw agent in under five minutes.

What You Will Build

An OpenClaw agent with live search across four platforms -- Google, Amazon, YouTube, and Walmart -- all from a single Scavio API key. The agent will be able to answer questions like "Find standing desks under $300 on Amazon with free shipping" or "Search YouTube for LangGraph tutorials from 2026" by calling the right skill automatically.

Prerequisites

• OpenClaw installed and a working agent project
• A Scavio API key -- sign up at the Scavio Dashboard (free tier includes 1,000 credits/month, no credit card required)
• The clawhub CLI -- comes bundled with OpenClaw

Step 1: Install the Skills

Scavio publishes four separate skills on ClawHub -- one per platform. Install only what your agent needs, or grab all four at once:

clawhub install scavio-google scavio-amazon scavio-youtube scavio-walmart

Each skill is self-contained. Installing scavio-google does not pull in Amazon or YouTube -- your agent only gets the tools you explicitly add.

Step 2: Set Your API Key

All four skills read your Scavio API key from a single environment variable. Set it once and every installed skill picks it up automatically:

export SCAVIO_API_KEY=sk_live_your_key_here

Your key starts with sk_live_ for production or sk_test_ for testing. Both work with all skills. Find your key in the API Keys section of the dashboard.

Step 3: Verify the Skills Are Available

List your installed skills to confirm everything is in place:

clawhub list

# Output includes:
# scavio-google   v1.0.0   Real-time Google Search via Scavio
# scavio-amazon   v1.0.0   Amazon product search via Scavio
# scavio-youtube  v1.0.0   YouTube video search via Scavio
# scavio-walmart  v1.0.0   Walmart product search via Scavio

What Each Skill Does

Each skill maps natural-language instructions from your agent to the right Scavio API call and returns structured JSON. Here is what each one covers:

scavio-google

Searches Google and returns organic results, knowledge graphs, People Also Ask, related searches, and featured snippets. Supports web, news, images, and maps search types.

// Structured response includes:
{
  "results": [
    { "title": "...", "url": "...", "content": "...", "position": 1 }
  ],
  "knowledge_graph": { "title": "...", "description": "..." },
  "people_also_ask": [{ "question": "...", "answer": "..." }],
  "related_searches": [{ "query": "..." }]
}

See the Google Search API reference for the full response schema and available parameters.

scavio-amazon

Searches Amazon products and returns titles, prices, ratings, review counts, Prime eligibility, and best seller status. Supports 12 Amazon marketplaces.

// Product result includes:
{
  "title": "Flexispot E7 Pro Standing Desk",
  "price": 499.99,
  "rating": 4.6,
  "reviews_count": 2847,
  "url": "https://www.amazon.com/dp/...",
  "best_seller": false
}

scavio-youtube

Searches YouTube and returns video titles, channel names, view counts, publish dates, durations, and video URLs. Raw YouTube data is returned so your agent has the full context to reason about.

scavio-walmart

Searches Walmart products with prices, ratings, fulfillment options (delivery, free shipping, pickup), and stock status. Useful for comparison shopping agents.

// Product result includes:
{
  "title": "KitchenAid Stand Mixer 5qt",
  "price": 279.00,
  "rating": 4.8,
  "out_of_stock": false,
  "fulfillment": {
    "free_shipping": true,
    "delivery": true,
    "pickup": false
  }
}

Example Agent Prompts

Once the skills are installed, your agent can handle natural-language requests that map directly to search queries:

# Google Search
"What are the most popular AI agent frameworks in 2026?"
"Search for news about the latest OpenAI announcements"

# Amazon
"Find noise-cancelling headphones under $150 on Amazon"
"Look up the top-rated espresso machines on Amazon.co.uk"

# YouTube
"Find recent LangGraph tutorials on YouTube"
"Search YouTube for reviews of the M4 MacBook Pro"

# Walmart
"Find coffee makers with free shipping at Walmart under $50"
"Compare standing desk options available for pickup at Walmart"

Credit Usage

Each skill call consumes credits from your Scavio account. The cost depends on the request mode:

The free tier gives you 1,000 credits/month -- enough for 1,000 light searches or 500 full searches. For agents that run frequently, the Project plan ($30/mo) gives you 7,000 credits.

Choosing Between Scavio Skills and langchain-scavio

If you are building with OpenClaw, the ClawHub skills are the right choice -- zero configuration, no Python wrappers, and the agent discovers available tools automatically.

If you are building with LangChain or LangGraph, use the langchain-scavio Python package instead. It ships 7 typed tool classes with async support, LLM-controllable parameters, and native LangGraph ToolNode compatibility. See our guide on adding web search to a LangChain agent for a full walkthrough.

Next Steps

• Browse Scavio skills on ClawHub -- install page with version history and changelogs
• OpenClaw Integration docs -- reference for all four skills and configuration options
• Google Search API reference -- full parameter list and response schema
• Amazon API reference -- marketplace support, ASIN lookups, and product detail fields
• Quickstart -- direct REST API usage without any SDK or skill layer