tavily
starsforks
How to install tavily?
safeai-lab-x/clawkeeper/tavily
$npx -y skills add SafeAI-Lab-X/ClawKeeper --skill tavilyTavily web search, content extraction, and research tools.
SKILL.md
| 1 | # Tavily Tools |
| 2 | |
| 3 | ## When to use which tool |
| 4 | |
| 5 | | Need | Tool | When | |
| 6 | | ---------------------------- | ---------------- | ------------------------------------------------------------- | |
| 7 | | Quick web search | `web_search` | Basic queries, no special options needed | |
| 8 | | Search with advanced options | `tavily_search` | Need depth, topic, domain filters, time ranges, or AI answers | |
| 9 | | Extract content from URLs | `tavily_extract` | Have specific URLs, need their content | |
| 10 | |
| 11 | ## web_search |
| 12 | |
| 13 | Tavily powers this automatically when selected as the search provider. Use for |
| 14 | straightforward queries where you don't need Tavily-specific options. |
| 15 | |
| 16 | | Parameter | Description | |
| 17 | | --------- | ------------------------ | |
| 18 | | `query` | Search query string | |
| 19 | | `count` | Number of results (1-20) | |
| 20 | |
| 21 | ## tavily_search |
| 22 | |
| 23 | Use when you need fine-grained control over search behavior. |
| 24 | |
| 25 | | Parameter | Description | |
| 26 | | ----------------- | --------------------------------------------------------------------- | |
| 27 | | `query` | Search query string (keep under 400 characters) | |
| 28 | | `search_depth` | `basic` (default, balanced) or `advanced` (highest relevance, slower) | |
| 29 | | `topic` | `general` (default), `news` (real-time updates), or `finance` | |
| 30 | | `max_results` | Number of results, 1-20 (default: 5) | |
| 31 | | `include_answer` | Include an AI-generated answer summary (default: false) | |
| 32 | | `time_range` | Filter by recency: `day`, `week`, `month`, or `year` | |
| 33 | | `include_domains` | Array of domains to restrict results to | |
| 34 | | `exclude_domains` | Array of domains to exclude from results | |
| 35 | |
| 36 | ### Search depth |
| 37 | |
| 38 | | Depth | Speed | Relevance | Best for | |
| 39 | | ---------- | ------ | --------- | -------------------------------------------- | |
| 40 | | `basic` | Faster | High | General-purpose queries (default) | |
| 41 | | `advanced` | Slower | Highest | Precision, specific facts, detailed research | |
| 42 | |
| 43 | ### Tips |
| 44 | |
| 45 | - **Keep queries under 400 characters** — think search query, not prompt. |
| 46 | - **Break complex queries into sub-queries** for better results. |
| 47 | - **Use `include_domains`** to focus on trusted sources. |
| 48 | - **Use `time_range`** for recent information (news, current events). |
| 49 | - **Use `include_answer`** when you need a quick synthesized answer. |
| 50 | |
| 51 | ## tavily_extract |
| 52 | |
| 53 | Use when you have specific URLs and need their content. Handles JavaScript-rendered |
| 54 | pages and returns clean markdown. Supports query-focused chunking for targeted |
| 55 | extraction. |
| 56 | |
| 57 | | Parameter | Description | |
| 58 | | ------------------- | ------------------------------------------------------------------ | |
| 59 | | `urls` | Array of URLs to extract (1-20 per request) | |
| 60 | | `query` | Rerank extracted chunks by relevance to this query | |
| 61 | | `extract_depth` | `basic` (default, fast) or `advanced` (for JS-heavy pages, tables) | |
| 62 | | `chunks_per_source` | Chunks per URL, 1-5 (requires `query`) | |
| 63 | | `include_images` | Include image URLs in results (default: false) | |
| 64 | |
| 65 | ### Extract depth |
| 66 | |
| 67 | | Depth | When to use | |
| 68 | | ---------- | ----------------------------------------------------------- | |
| 69 | | `basic` | Simple pages — try this first | |
| 70 | | `advanced` | JS-rendered SPAs, dynamic content, tables, embedded content | |
| 71 | |
| 72 | ### Tips |
| 73 | |
| 74 | - **Max 20 URLs per request** — batch larger lists into multiple calls. |
| 75 | - **Use `query` + `chunks_per_source`** to get only relevant content instead of full pages. |
| 76 | - **Try `basic` first**, fall back to `advanced` if content is missing or incomplete. |
| 77 | - If `tavily_search` results already contain the snippets you need, skip the extract step. |
| 78 | |
| 79 | ## Choosing the right workflow |
| 80 | |
| 81 | Follow this escalation pattern — start simple, escalate only when needed: |
| 82 | |
| 83 | 1. **`web_search`** — Quick lookup, no special options needed. |
| 84 | 2. **`tavily_search`** — Need depth control, topic filtering, domain filters, time ranges, or AI answers. |
| 85 | 3. **`tavily_extract`** — Have specific URLs, need their full content or targeted chunks. |
| 86 | |
| 87 | Combine search + extract when you need to find pages first, then get their full content. |
People who liked this also viewed
6 picksType
- find-skillsHelps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skSkillsJul 2026
- grill-meA relentless interview to sharpen a plan or design.SkillsJul 2026
- vercel-react-best-practicesVercel's official collection of agent skillsSkillsJul 2026
- improve-codebase-architectureScan a codebase for deepening opportunities, present them as a visual HTML report, then grill through whichever one you pick.SkillsJul 2026
- setup-matt-pocock-skillsConfigure this repo for the engineering skills — set up its issue tracker, triage label vocabulary, and domain doc layout. Run once before first use of the other engineering skills.SkillsJul 2026
- triageMove issues and external PRs through a state machine of triage roles — categorise, verify, grill if needed, and write agent-ready briefs.SkillsJul 2026
