# Tavily Research

> Produce cited multi-source research. Use when the user wants a deep report, comparison, market analysis, or literature review, or says "research", "investigate", "analyze in depth", "compare X vs Y", or "what does the market look like for". For a quick fact, use tavily-search instead.

- Skill: `tavily-ai/tavily-research-3` (Agent Skill)
- Install (CLI): `npx skillmds@latest add tavily-ai/tavily-research-3`
- Raw SKILL.md: https://api.skillmd.com/api/skills/tavily-ai/tavily-research-3/raw
- Safety review: pending
- Works with: Claude Code, Claude.ai, OpenAI Codex
- Category: Docs & Writing
- Author: tavily-ai (https://skillmd.com/u/tavily-ai)
- Updated: 2026-09-22
- Page: https://skillmd.com/skills/tavily-ai/tavily-research-3

---


# Tavily Research

Synthesize a grounded report from multiple sources, with citations. This is slower than search.

## When

- The user needs comparison, market context, or a literature-style brief
- Quick search snippets are not enough
- Step 5: search → extract → map → crawl → **research**

Do not use this skill for a simple lookup (search) or for pulling one site’s docs (crawl).

## How

- Clarify scope, freshness, geography, and output shape only when they are ambiguous.
- Wait for the research run to finish. Preserve source URLs and distinguish sourced facts from your analysis.
- Prefer research when the deliverable is the report. Prefer search when the user wants a short answer now.

## Execute

If `tavily_research` is callable (the name may be prefixed), call it. Do not install the CLI and do not run `tvly` for this. In a coding-agent terminal that is already using `tvly`, pass `--client-name "cursor plugin"` and take flags from `tvly research --help`. Never ask for an API key in chat.