Skip to content

Claude/streamline search providers giw xu#9

Open
Karthik777 wants to merge 2 commits into
mainfrom
claude/streamline-search-providers-GiwXU
Open

Claude/streamline search providers giw xu#9
Karthik777 wants to merge 2 commits into
mainfrom
claude/streamline-search-providers-GiwXU

Conversation

@Karthik777

Copy link
Copy Markdown
Contributor

No description provided.

claude added 2 commits May 17, 2026 00:12
Cut webba down to a light, efficient core:

- Providers: keep searxng, perplexity, fastcdp (browser SERP scrape via CDP);
  remove ddg, google_scrape, serper, tavily, exa, brave
- Remove QuotaManager and free-tier quota tracking
- Replace SemanticSearchCache (model2vec + litesearch) with a diskcache-backed
  literal-query TTL cache
- Replace the ProviderHealth cooldown system with Perplexity HTTP error-code
  handling (401/403/404/400 fatal; 429/5xx retried once)
- Replace the hand-rolled fetch cascade with scrapling (fast HTTP -> stealth
  browser -> dynamic browser)
- Tune SearXNG engines (drop noisy ones, add YouTube); add yt-dlp transcript
  extraction for YouTube URLs
- Add chrome_debug_setup() for isolated-profile, login-gated search
- Remove the Hermes Agent plugin, skill.py, and entry points

https://claude.ai/code/session_01F6MxPQNN1s88ttzfnmHzxW
Drop webba/cache.py entirely. search.py now exposes a shared diskcache
store (LRU, 500 MB cap) and a small @cache decorator built on
memoize_stampede; _search_cached is the decorated search core and
search(cache=False) bypasses it. purge_cache() wipes the store.

https://claude.ai/code/session_01F6MxPQNN1s88ttzfnmHzxW
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.

2 participants