| name | web-fetch |
| description | Read supplied public URL(s) into readable content, metadata, or links; optionally select Exa, Tavily, or Firecrawl. |
Web Fetch
Read supplied URL(s) through the configured provider selected by registry/web-fetch-providers.yaml. For ambiguous web intent, apply workflows/web-capability-routing.md. Accept portable controls for URLs, output format, main-content extraction, wait time, and content limits. Report the actual provider and any fallback.
- Prefer Firecrawl for rendered or difficult pages.
- Prefer Exa for Exa result IDs, highlights, summaries, or related content.
- Prefer Tavily for simple URL extraction.
- Do not silently substitute a user-requested provider.
Run
python3 <installed-arkspace-path>/scripts/arkspace.py web fetch --provider firecrawl https://example.com --output json
For setup or a failed provider check, use provider-manager; never ask users to edit provider files or expose keys.