Quickstart

The core loop: search → read → store → query. Every command below works immediately after go install.

1. Search the web

webx search "concurrent map writes golang panic"

Fused metasearch over 15 providers — weighted RRF + BM25 blending + domain priors. score is normalized 0–1, near-duplicates are dropped (deduped in output). Check provider health anytime with webx doctor.

2. Read a page

webx scrape https://go.dev/doc/effective_go

Clean markdown + metadata through the readability→trafilatura→full-page cascade. For JS-heavy pages:

webx scrape https://spa.example.com --auto-render   # escalates http → browser automatically
webx scrape https://go.dev --fit "http server"      # BM25-filtered markdown — the token-saver

3. Search and read in one hop

webx ask "how does htmx hx-swap work"

Search → scrape top results → citation-numbered excerpts under a token budget. Add --llm for a synthesized cited answer (needs WEBX_LLM_*).

4. Build a private index

webx index go.dev
webx query "error handling patterns"                # lexical FTS5, offline
webx query --semantic "graceful shutdown"           # + embeddings (WEBX_EMBED_MODEL)

index crawls the sitemap (or BFS) into ~/.webx/index.db. --semantic fuses bm25 with chunk-level embedding cosine — Exa-style neural search over your corpus.

5. Serve it

webx serve                                          # HTTP API on :8080
curl -X POST localhost:8080/scrape \
  -H 'content-type: application/json' \
  -d '{"url":"https://go.dev/doc/effective_go"}'

Firecrawl /v1//v2 routes are built in — point a Firecrawl SDK at the same base URL and it works. Full surface: HTTP API.

Next steps

• Fetch tiers & rendering — when http stops being enough
• Agents, MCP & skills — wire webx into Claude Code, Cursor, Devin
• Command map — all 30 commands with flags