Tavily Research - Setup Guide¶

Source: tavily-ai/skills (14,100+ installs) Category: Web Search / Research Quality Tier: 🟢 Production

AI-powered deep research agent that gathers web sources, synthesizes findings, and produces structured reports with explicit citations. Part of the Tavily workflow pipeline: search → extract → map → crawl → research. Takes 30-120 seconds for comprehensive analysis.

Installation¶

# Install Tavily CLI (if not already)
curl -fsSL https://cli.tavily.com/install.sh | bash

# Login
tvly login

Prerequisites¶

Requirement	Details
Tavily API key	Sign up at tavily.com
tvly CLI	Installed via command above
Research credits	Research consumes more credits than search - Pro model requires paid plan

Key Capabilities¶

When to Use¶

Comprehensive multi-source analysis needed
Comparison between products, frameworks, or strategies
Market reports or industry landscaping
Literature reviews with citations
When quick searches aren't enough - you need synthesis

When NOT to Use¶

Quick fact-finding → use tavily-search instead
Single-page extraction → use tavily-extract
Known URL analysis → use firecrawl scrape

Quick Start¶

# Basic research (waits for completion)
tvly research "competitive landscape of AI code assistants in 2026"

# Pro model for comprehensive analysis
tvly research "electric vehicle battery technology market analysis" --model pro

# Stream results in real-time
tvly research "AI agent frameworks comparison LangChain vs CrewAI vs AutoGen" --stream

# Save report to file
tvly research "fintech payment processing trends 2026" --model pro -o fintech-report.md

# Research with specific depth
tvly research "quantum computing commercial applications" --depth advanced

Options Reference¶

Option	Description
`--model`	`basic` (default) or `pro` (comprehensive, requires paid plan)
`--stream`	Stream results as they're generated
`--depth`	Research depth: `basic` or `advanced`
`-o, --output <path>`	Save report to file (markdown)
`--max-results`	Max sources to analyze

Hermes Integration¶

For CorpusIQ growth operations, Tavily Research is ideal for:

Market Intelligence¶

tvly research "AI-powered business intelligence tools market landscape 2026" \
  --model pro -o ~/corpusiq-brain/research/ai-bi-landscape.md

Competitive Analysis¶

tvly research "Compare CorpusIQ alternatives: features, pricing, target audience, strengths, weaknesses" \
  --stream -o ~/corpusiq-brain/research/competitive-analysis.md

Content Research¶

tvly research "What problems do ecommerce operators face with data consolidation across Shopify, Meta Ads, and Google Analytics" \
  --model pro -o ~/corpusiq-brain/content/operator-pain-points.md

Tips¶

Research runs take 30-120 seconds - use --stream for progress visibility
Pro model provides significantly more comprehensive analysis but requires paid Tavily plan
Save reports to your knowledge base (GBrain) for persistent access
Combine with tavily-search for initial exploration, then tavily-research for final synthesis

Troubleshooting¶

Issue	Solution
Pro model unavailable	Upgrade Tavily plan at tavily.com/pricing
Research timeout	Simplify the query or reduce depth
Empty citations	Broaden the research topic or use `--depth advanced`
Report too long	Use `-o` to save to file instead of stdout

Integration with CorpusIQ¶

CorpusIQ is the connector layer that lets AI assistants read the systems a business already runs on - accounting, payments, CRM, email, files and analytics - and answer with the source named. It is read-only by design: it retrieves data and never writes to, or acts in, a connected system, which is what makes it safe to point at live accounts.

This page is part of the CorpusIQ directory, the reference we maintain for the MCP server and agent-skill ecosystem. Run alongside CorpusIQ, a third-party server like this one can be combined with a company's own business data in a single answer.

See what CorpusIQ connects