Update to Tavily

{
  "name": "tavily-best-practices",
  "description": "Build production-ready Tavily integrations with best practices baked in. Reference documentation for developers using coding assistants (Claude Code, Cursor, etc.) to implement web search, content extraction, crawling, and research in agentic workflows, RAG systems, or autonomous agents.",
  "included_files": [
    {
      "relative_path": "references/crawl.md",
      "size_in_bytes": 10204
    },
    {
      "relative_path": "references/extract.md",
      "size_in_bytes": 7381
    },
    {
      "relative_path": "references/integrations.md",
      "size_in_bytes": 17980
    },
    {
      "relative_path": "references/research.md",
      "size_in_bytes": 10090
    },
    {
      "relative_path": "references/sdk.md",
      "size_in_bytes": 8353
    },
    {
      "relative_path": "references/search.md",
      "size_in_bytes": 12088
    }
  ],
  "skill_md_contents": "---\nname: tavily-best-practices\ndescription: \"Build production-ready Tavily integrations with best practices baked in. Reference documentation for developers using coding assistants (Claude Code, Cursor, etc.) to implement web search, content extraction, crawling, and research in agentic workflows, RAG systems, or autonomous agents.\"\n---\n\n# Tavily\n\nTavily is a search API designed for LLMs, enabling AI applications to access real-time web data.\n\n## Installation\n\n**Python:**\n```bash\npip install tavily-python\n```\n\n**JavaScript:**\n```bash\nnpm install @tavily/core\n```\n\nSee **[references/sdk.md](references/sdk.md)** for complete SDK reference.\n\n## Client Initialization\n\n```python\nfrom tavily import TavilyClient\n\n# Uses TAVILY_API_KEY env var (recommended)\nclient = TavilyClient()\n\n#With project tracking (for usage organization)\nclient = TavilyClient(project_id=\"your-project-id\")\n\n# Async client for parallel queries\nfrom tavily import AsyncTavilyClient\nasync_client = AsyncTavilyClient()\n```\n\n## Choosing the Right Method\n\n**For custom agents/workflows:**\n\n| Need | Method |\n|------|--------|\n| Web search results | `search()` |\n| Content from specific URLs | `extract()` |\n| Content from entire site | `crawl()` |\n| URL discovery from site | `map()` |\n\n**For out-of-the-box research:**\n\n| Need | Method |\n|------|--------|\n| End-to-end research with AI synthesis | `research()` |\n\n## Quick Reference\n\n### search() - Web Search\n\n```python\nresponse = client.search(\n    query=\"quantum computing breakthroughs\",  # Keep under 400 chars\n    max_results=10,\n    search_depth=\"advanced\"\n)\nprint(response)\n```\nKey parameters: `query`, `max_results`, `search_depth` (ultra-fast/fast/basic/advanced), `include_domains`, `exclude_domains`, `time_range`\n\nSee **[references/search.md](references/search.md)** for complete search reference.\n\n### extract() - URL Content Extraction\n\n```python\n# Simple one-step extraction\nresponse = client.extract(\n    urls=[\"https://docs.example.com\"],\n    extract_depth=\"advanced\"\n)\nprint(response)\n```\nKey parameters: `urls` (max 20), `extract_depth`, `query`, `chunks_per_source` (1-5)\n\nSee **[references/extract.md](references/extract.md)** for complete extract reference.\n\n### crawl() - Site-Wide Extraction\n\n```python\nresponse = client.crawl(\n    url=\"https://docs.example.com\",\n    instructions=\"Find API documentation pages\",  # Semantic focus\n    extract_depth=\"advanced\"\n)\nprint(response)\n```\nKey parameters: `url`, `max_depth`, `max_breadth`, `limit`, `instructions`, `chunks_per_source`, `select_paths`, `exclude_paths`\n\nSee **[references/crawl.md](references/crawl.md)** for complete crawl reference.\n\n### map() - URL Discovery\n\n```python\nresponse = client.map(\n    url=\"https://docs.example.com\"\n)\nprint(response)\n```\n\n### research() - AI-Powered Research\n\n```python\nimport time\n\n# For comprehensive multi-topic research\nresult = client.research(\n    input=\"Analyze competitive landscape for X in SMB market\",\n    model=\"pro\"  # or \"mini\" for focused queries, \"auto\" when unsure\n)\nrequest_id = result[\"request_id\"]\n\n# Poll until completed\nresponse = client.get_research(request_id)\nwhile response[\"status\"] not in [\"completed\", \"failed\"]:\n    time.sleep(10)\n    response = client.get_research(request_id)\n\nprint(response[\"content\"])  # The research report\n```\n\nKey parameters: `input`, `model` (\"mini\"/\"pro\"/\"auto\"), `stream`, `output_schema`, `citation_format`\n\nSee **[references/research.md](references/research.md)** for complete research reference.\n\n## Detailed Guides\n\nFor complete parameters, response fields, patterns, and examples:\n\n- **[references/sdk.md](references/sdk.md)** - Python & JavaScript SDK reference, async patterns, Hybrid RAG\n- **[references/search.md](references/search.md)** - Query optimization, search depth selection, domain filtering, async patterns, post-filtering\n- **[references/extract.md](references/extract.md)** - One-step vs two-step extraction, query/chunks for targeting, advanced mode\n- **[references/crawl.md](references/crawl.md)** - Crawl vs Map, instructions for semantic focus, use cases, Map-then-Extract pattern\n- **[references/research.md](references/research.md)** - Prompting best practices, model selection, streaming, structured output schemas\n- **[references/integrations.md](references/integrations.md)** - LangChain, LlamaIndex, CrewAI, Vercel AI SDK, and framework integrations\n"
}