Main Web Search entry point for public evidence, APIs, versions and comparisons. Normally omit engines: engine_pool chooses which sources to search; ranking changes only final engine scoring weights; complexity controls budget, query variants and depth. Optional engine_weights override scoring, never engine selection. Enable community only when recent developer/community voices on X are relevant (default false); use x_search directly for X-only account/thread searches. Returns scoreVersion, ranked evidence, enginesUsed, effectiveWeights, communityUsed and warnings. Consensus-v2 scores are not probabilities; legacy min_score thresholds need recalibration. Current availability is reported by runtime capabilities, not guaranteed by a preset. Use fetch_page when snippets are insufficient. Optional live status: search-boost://capabilities Resource.

Read a known http(s) URL without searching again. Fetches the origin first, cleans HTML, and uses same-route curl compatibility fallback and Jina Reader backup when needed. Strips CSS/JS/ad chrome. Use focus for matching paragraphs, or omit it for the readable body. A focus miss is not proof the page lacks the answer: retry without focus. Proxies handle destination DNS; direct routes use local networking. Five proxy connection failures permit direct fallback. HTTP(S), TLS, cancellation and response-size limits still apply. Not an authenticated browser.

Find X/Twitter posts, inspect an account, or retrieve available thread material. Use keyword/semantic with query, user with username, or thread with post_id. Credential-free retrieval is available; configured X authentication can improve coverage. Author/date filters apply across retrieval paths; candidates whose required metadata cannot be verified are omitted. Results may be incomplete, delayed or empty; a post sample does not establish platform-wide sentiment and a thread result is not guaranteed to contain the full conversation.

Inspect the current search layer with layer=show (default, no change). layer=free or api persists a new default: only change it when authorized. free = free — compatibility default engine_pool=free (keyless). api = api — compatibility default engine_pool=hybrid (keyless plus configured APIs). For a single search, use fused_search.engine_pool instead; legacy api maps to hybrid, not the strict api pool.

Read-only diagnostics for failed or empty searches: cache hits/misses, tier counts, engine availability, and recent activity. Call with no arguments. Inspect tool warnings too; an empty result alone does not imply missing credentials or justify changing configuration.

Jev-driven keyword-target research. Supply tasks with context and targets (id, keywords, question), or legacy questions; at most 6 tasks and 12 targets total. Each round has at most three batched Jev calls: plan, evidence scoring, coverage/continuation. Returns only Jev-approved URLs, titles and extractive descriptions, not rejected or unassessed candidates. Approval is a model judgement, not verified truth. Follow nextCursor using cursor (without tasks/questions) to read more, without search or Jev calls. page_size defaults to 20 (max 50); there is no total approved-result count cap, but search time/request budgets remain finite. Pagination is not exhaustive search: inspect coverageComplete and warnings. Results are held in this server process for up to 30 minutes / 32 recent runs and are lost on restart or eviction. Relative today/yesterday constraints use UTC; unknown event dates do not qualify. Requires Jev credentials; otherwise returns not_configured without network requests. Task text and necessary fragments go to the configured Jev service (default TypeSafe), never engine credentials. Prefer fused_search for a precise lookup, fetch_page for a known URL and x_search for X-specific retrieval.