Skip to content

overhaul web search and fetch tooling#117

Merged
jbj338033 merged 4 commits into
mainfrom
worktree-improve-search
Jul 14, 2026
Merged

overhaul web search and fetch tooling#117
jbj338033 merged 4 commits into
mainfrom
worktree-improve-search

Conversation

@jbj338033

Copy link
Copy Markdown
Collaborator

Summary

WebSearch returned empty results ~84% of the time because the default no-key DuckDuckGo backend is bot-blocked in 2026. This reworks WebSearch to be API-first (Tavily/Brave/SearXNG) with reliable-provider fallback, treats an empty result as a failure instead of "no results exist", and returns actionable setup guidance (plus a new /search command) when nothing is configured. WebFetch is overhauled for lossless, high-quality extraction: dom_smoothie main-content readability, encoding_rs charset decoding, content-type routing (JSON pretty-print, pdf_oxide PDF text, typed binary notices), offset/max_length pagination, BM25 query-guided passages, page metadata + summary, and an automatic headless-chromium render fallback for JavaScript pages.

Verification

  • cargo fmt --all --check — passed
  • cargo clippy --workspace --all-targets -- -D warnings — passed (0 warnings)
  • cargo test --workspace — passed (717 tests)

@jbj338033
jbj338033 merged commit 016d4d8 into main Jul 14, 2026
4 of 5 checks passed
@jbj338033
jbj338033 deleted the worktree-improve-search branch July 14, 2026 00:28
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant