Web Research (Native - No API Keys)

Search the web and fetch content using local tools. NO API KEYS REQUIRED.

Web Search with ddgr (DuckDuckGo)

JSON Output (Best for Parsing)

ddgr -n 5 --json "your search query"

Returns structured JSON:

[
  {"title": "Page Title", "url": "https://...", "abstract": "Description..."},
  ...
]

Human Readable

ddgr -n 5 "your search query"

Advanced Options

# More results
ddgr -n 10 --json "query"

# Site-specific search
ddgr -n 5 --json "site:github.com ollama setup"

# Time-limited (past year)
ddgr -n 5 --json --time y "latest news topic"

Fetch URL Content with w3m

Basic Fetch

w3m -dump "https://example.com" | head -c 50000

Fetch and Pipe

w3m -dump "https://example.com" | grep -i "keyword"

Research Workflow

1. •

   Search - Find relevant URLs:

   bash

   ddgr -n 5 --json "topic of interest"
   

2. •

   Fetch - Get content from best result:

   bash

   w3m -dump "https://best-result-url.com" | head -c 30000
   

3. •

   Summarize - Process and present findings

Scripts (Shortcuts)

# Search script
./scripts/websearch.sh "query"

# Fetch script
./scripts/webfetch.sh "url"

Tips

• •Use --json for machine-parseable output
• •Pipe through jq for JSON processing: ddgr -n 5 --json "query" | jq '.[0].url'
• •Limit fetch size with head -c BYTES to avoid overwhelming context
• •For JavaScript-heavy sites, use the browser tool instead

When to Use What