Search Skill
Power your ClawMesh agents with real-time web search capabilities. The Search Skill integrates Tavily AI search for accurate, context-aware information retrieval from across the internet.
- ›Real-time web search with Tavily AI API
- ›Deep search mode for comprehensive research
- ›Content extraction and structured data retrieval
- ›Multi-source search aggregation
Get Started
Enable real-time search for your agents
Install the Search Skill and give your AI agents internet search capabilities.
What it does
The Search Skill gives your ClawMesh agent the ability to search the web in real-time, extract content from any website, and retrieve accurate, up-to-date information. Unlike traditional keyword-based search, the Tavily integration understands search intent and returns contextually relevant results.
This skill transforms your agent from a static knowledge system into a dynamic research assistant that can answer questions about current events, verify facts, monitor competitors, and gather information from across the web on any topic.
Why Tavily?
Tavily is an AI-native search API designed specifically for large language models and AI agents. Unlike generic search engines, Tavily returns structured, LLM-friendly results with relevance scores and source attribution. The API is optimized for deep research tasks, not just surface-level keyword matching.
Tavily's deep search mode is particularly powerful for agentic use cases. It performs multi-step research, following logical chains between topics and synthesizing information from multiple sources into coherent, cited responses. This makes it ideal for the Search Skill in ClawMesh agents conducting research workflows.
Key features
- Real-time web search with AI-optimized relevance ranking
- Deep search mode for comprehensive multi-source research
- Structured JSON responses optimized for AI consumption
- Content extraction with article summaries and key findings
- Relevance scoring for every result
- Source attribution and citation support
- Image search and thumbnail extraction
- Academic and news source filtering
Configuration example
The Search Skill uses YAML configuration to define search behavior. Here's a typical configuration for the Tavily provider:
api_key: ${TAVILY_API_KEY}
search_mode: deep # or 'comprehensive' for research
max_results: 10
include_answer: true
include_images: false
include_raw_content: false
days: 30 # recency filter for news
Environment setup
Before using the Search Skill, you need a Tavily API key. Sign up at tavily.com to get your free API key. The free tier includes 1000 searches per month. For production use, consider the paid plans with higher rate limits.
Set the TAVILY_API_KEY environment variable in your ClawMesh configuration. Never hardcode API keys in configuration files that might be committed to version control.
Use cases
Competitive research: Monitor competitor websites, press releases, and industry news. Your agent can compile weekly competitive intelligence reports automatically.
Real-time fact verification: Before responding to user queries, agents can verify claims against current sources, reducing hallucinations and improving accuracy.
Market monitoring: Track specific keywords, brands, or topics across the web. Get alerts when new content matches your criteria.
Content aggregation: Pull together information from multiple sources on any topic, synthesizing perspectives into a comprehensive summary.
Academic research: Filter searches to academic sources and papers for literature reviews and research compilation.
Installation
Install the Search Skill from the ClawMesh dashboard, or use the CLI: clawmesh skills install search
After installation, configure your Tavily API key in the skill settings. If you don't have a key, visit tavily.com to sign up. The free tier provides 1000 searches per month.
Related guides
Q&A
What is Tavily and why is it used for AI search?
Tavily is an AI-native search API designed specifically for LLM and AI agent applications. Unlike traditional search APIs that return keyword-matched results, Tavily uses AI to understand search intent and return contextually relevant, structured results optimized for AI consumption. It's built by researchers who understand how AI systems process and use search results.
How accurate is the search results?
Tavily's search accuracy is among the best for AI-native search APIs. It uses multiple relevance signals and AI-powered ranking to return high-quality results. The API also provides relevance scores for each result, allowing agents to filter or prioritize based on confidence levels.
What's the difference between basic and deep search modes?
Basic search mode returns quick, relevant results suitable for simple queries. Deep search mode performs multi-step research, following logical connections between topics and synthesizing information from multiple sources. For research-intensive agent workflows, deep search mode produces significantly better results at the cost of slightly higher API usage.
Are there rate limits?
Rate limits depend on your Tavily plan. Free tier includes 1000 searches per month. Paid plans offer higher limits suitable for production workloads. The Search Skill includes caching to minimize redundant API calls and maximize your search budget.