Tavily Web Seeker

Enhanced Tavily search with intelligent intent recognition, source preferences, critical source validation, and offline report generation.

Enhanced web search skill built on Tavily API with intelligent search strategies, user source preferences, and critical source handling.

Features

1. Intent Recognition - Auto-classify queries (factual, news, research, comparison, how-to, academic)
2. Source Preferences - Topic-based domain routing (tech→GitHub, news→Reuters, finance→WSJ)
3. Search Operators - Support OR, site:, -, exact phrase, time filters
4. Critical Source Check - Validate results against preferred sources, return not-found conclusions
5. Multi-API Key Rotation - Auto-failover on rate limits (429)
6. Offline Report - Markdown export with source coverage analysis

Tools

Tool	Description
tavily_search	Standard Tavily search
web_seeker	Enhanced search with all features

Configuration

API Keys

json

{
  "apiKey": { "key1": "tvly-xxx1", "key2": "tvly-xxx2" }
}

Or env: TAVILY_API_KEY, TAVILY_API_KEY_2

Source Preferences

javascript

// User preferences
const PREFERENCES = {
  default_sites: ["wikipedia.org", "github.com"],
  preferred_topics: {
    tech: ["wikipedia.org", "github.com", "stackoverflow.com"],
    news: ["reuters.com", "apnews.com", "bbc.com"],
    finance: ["wsj.com", "bloomberg.com", "ft.com"],
    academic: ["scholar.google.com", "arxiv.org"]
  },
  critical_sources: {
    "官方": ["gov.cn", "moe.gov.cn"],
    "财经": ["wsj.com", "bloomberg.com"],
    "技术": ["github.com", "stackoverflow.com"]
  }
};

web_seeker Parameters

Parameter	Type	Default	Description
query	string	-	Search query
mode	string	"quick"	quick/research/compare
max_results	number	5	Results 1-20
domains	array	auto	Preferred domains
time_range	string	auto	day/week/month/year
operators	string	-	Raw operators
strict_mode	boolean	false	Require preferred sources only
export_doc	boolean	false	Generate report

Response Format

json

{
  "original_query": "Iran oil news",
  "intent": "news",
  "domains_used": ["reuters.com", "bbc.com"],
  "search_operator_applied": "site:reuters.com OR site:bbc.com",
  "results_found": true,
  "source_coverage": { "preferred": 3, "total": 10 },
  "results": [...],
  "summary": "..."
}

Not Found Response

json

{
  "results_found": false,
  "not_found_conclusion": "在指定的核心信源中未搜索到...",
  "suggestions": ["扩展搜索词", "放宽信源限制", "手动访问目标网站"]
}

Usage

bash

# Basic search
node seeker.js "AI news"

# Research mode with report
node seeker.js "machine learning" --mode research --export

# Strict mode (require preferred sources)
node seeker.js "python tutorial" --strict