Wigolo
Local-first web intelligence MCP server — search the web, fetch pages as markdown, crawl sites, extract structured data, research topics, and monitor pages for changes. All operations run locally with a built-in knowledge cache.
When to use
- Searching the web for information
- Fetching web pages as clean markdown
- Crawling sites to build a local knowledge base
- Extracting structured data from pages
- Conducting multi-step research on a topic
- Monitoring pages for changes over time
- Finding similar content to a known page
- Diffing a page against its cached copy
Core tools
| Tool | Purpose |
|---|---|
search <query> |
Search the web |
fetch <url> |
Fetch a page as markdown |
crawl <url> |
Crawl a site into the cache |
extract <url> |
Extract structured data from a page |
research <question> |
Multi-step research brief |
agent <prompt> |
Autonomous data gathering |
find-similar <url> |
Find related pages |
diff <url> |
Diff a page against cached copy |
watch <subcommand> |
Manage change-watch jobs |
cache <subcommand> |
Query the local knowledge cache |
MCP integration
Start the MCP server:
wigolo # stdio transport (default)
wigolo serve --port N # HTTP daemon
Setup and diagnostics:
wigolo init # First-time setup
wigolo doctor --fix # Diagnose and repair
wigolo health --json # Health check
wigolo verify --json # End-to-end smoke check
Related skills
- agent-reach: Multi-platform internet research (13 platforms)
- deep-research: Multi-source deep research with firecrawl + exa
- search-first: Research-before-coding workflow
- research-ops: Evidence-first current-state research
- firecrawl: Web scraping and crawling (alternative backend)
- scrapling: Web scraping MCP (Python, alternative backend)