@staticn0va/wigolo 0.1.0 → 0.1.1

package/dist/tools/search.js

@@ -1,160 +1,15 @@
-import { deduplicateResults } from '../search/dedup.js';
-import { decomposeQuery } from '../search/query.js';
-import { validateLinks } from '../search/validator.js';
-import { rerankResults } from '../search/rerank.js';
-import { applyAllFilters } from '../search/filters.js';
-import { extractContent } from '../extraction/pipeline.js';
-import { cacheSearchResults, getCachedSearchResults } from '../cache/store.js';
-import { getConfig } from '../config.js';
-import { createLogger } from '../logger.js';
-const log = createLogger('search');
-const DEFAULT_MAX_RESULTS = 5;
-const MAX_RESULTS_CAP = 20;
-const DEFAULT_CONTENT_MAX_CHARS = 30000;
-const DEFAULT_MAX_TOTAL_CHARS = 50000;
-export async function handleSearch(input, engines, router) {
-    const start = Date.now();
-    const config = getConfig();
-    const maxResults = Math.min(input.max_results ?? DEFAULT_MAX_RESULTS, MAX_RESULTS_CAP);
-    const includeContent = input.include_content ?? true;
-    const contentMaxChars = input.content_max_chars ?? DEFAULT_CONTENT_MAX_CHARS;
-    const maxTotalChars = input.max_total_chars ?? DEFAULT_MAX_TOTAL_CHARS;
-    const totalTimeoutMs = config.searchTotalTimeoutMs;
-    const fetchTimeoutMs = config.searchFetchTimeoutMs;
-    const cached = getCachedSearchResults(input.query);
-    if (cached && !includeContent) {
-        log.info('serving search results from cache', { query: input.query });
-        return {
-            results: cached.results.slice(0, maxResults),
-            query: input.query,
-            engines_used: cached.engines_used,
-            total_time_ms: Date.now() - start,
-        };
-    }
-    let activeEngines = engines;
-    if (input.search_engines && input.search_engines.length > 0) {
-        activeEngines = engines.filter(e => input.search_engines.includes(e.name));
-        if (activeEngines.length === 0) {
-            log.warn('no engines matched search_engines filter, using all', { requested: input.search_engines });
-            activeEngines = engines;
-        }
-    }
-    const subQueries = decomposeQuery(input.query);
-    log.debug('query decomposition', { original: input.query, parts: subQueries.length });
-    const allRaw = [];
-    const enginesUsed = new Set();
-    const errors = [];
-    // Increase overfetch when domain filters are active (more results will be filtered out)
-    const hasFilterAttrition = !!(input.include_domains?.length || input.exclude_domains?.length);
-    const overfetchFactor = hasFilterAttrition ? 3 : 2;
-    const searchPromises = activeEngines.flatMap(engine => subQueries.map(async (query) => {
-        try {
-            const results = await engine.search(query, {
-                maxResults: maxResults * overfetchFactor,
-                timeRange: input.time_range,
-                language: input.language,
-                includeDomains: input.include_domains,
-                excludeDomains: input.exclude_domains,
-                fromDate: input.from_date,
-                toDate: input.to_date,
-                category: input.category,
-            });
-            for (const r of results) {
-                allRaw.push(r);
-                enginesUsed.add(engine.name);
-            }
-        }
-        catch (err) {
-            const msg = err instanceof Error ? err.message : String(err);
-            log.warn('engine search failed', { engine: engine.name, query, error: msg });
-            errors.push(`${engine.name}: ${msg}`);
-        }
-    }));
-    await Promise.allSettled(searchPromises);
-    if (allRaw.length === 0) {
-        return {
-            results: [],
-            query: input.query,
-            engines_used: [...enginesUsed],
-            total_time_ms: Date.now() - start,
-            error: errors.length > 0 ? errors.join('; ') : 'No results found',
-        };
-    }
-    let merged = deduplicateResults(allRaw);
-    // Post-filter: domain + date + category (Slice 7) — runs before rerank so
-    // the reranker only scores results that pass filters (Slice 9 interaction).
-    merged = applyAllFilters(merged, {
-        includeDomains: input.include_domains,
-        excludeDomains: input.exclude_domains,
-        fromDate: input.from_date,
-        toDate: input.to_date,
-        category: input.category,
-    });
-    merged = await rerankResults(input.query, merged);
-    merged = await validateLinks(merged);
-    merged = merged.slice(0, maxResults);
-    const results = merged.map(m => ({
-        title: m.title,
-        url: m.url,
-        snippet: m.snippet,
-        relevance_score: m.relevance_score,
-    }));
-    if (includeContent && results.length > 0) {
-        await fetchContentForResults(results, router, {
-            contentMaxChars,
-            maxTotalChars,
-            fetchTimeoutMs,
-            totalDeadline: start + totalTimeoutMs,
-        });
-    }
-    try {
-        cacheSearchResults(input.query, results, [...enginesUsed]);
-    }
-    catch (err) {
-        log.warn('failed to cache search results', { error: String(err) });
-    }
-    return {
-        results,
-        query: input.query,
-        engines_used: [...enginesUsed],
-        total_time_ms: Date.now() - start,
-    };
-}
-// v1: sequential fetch for correct budget enforcement. v2: parallel fetch then apply budget in relevance order.
-async function fetchContentForResults(results, router, ctx) {
-    let totalCharsUsed = 0;
-    for (const result of results) {
-        if (Date.now() >= ctx.totalDeadline) {
-            result.fetch_failed = 'total_timeout';
-            continue;
-        }
-        if (totalCharsUsed >= ctx.maxTotalChars) {
-            result.content_truncated = true;
-            continue;
-        }
-        try {
-            const raw = await Promise.race([
-                router.fetch(result.url, { renderJs: 'auto' }),
-                new Promise((_, reject) => setTimeout(() => reject(new Error('timeout')), ctx.fetchTimeoutMs)),
-            ]);
-            const extraction = await extractContent(raw.html, raw.finalUrl, {
-                maxChars: ctx.contentMaxChars,
-                contentType: raw.contentType,
-            });
-            let content = extraction.markdown;
-            const remaining = ctx.maxTotalChars - totalCharsUsed;
-            if (content.length > remaining) {
-                content = content.slice(0, remaining);
-                result.content_truncated = true;
-            }
-            totalCharsUsed += content.length;
-            result.markdown_content = content;
-        }
-        catch (err) {
-            const msg = err instanceof Error ? err.message : String(err);
-            log.debug('content fetch failed', { url: result.url, error: msg });
-            result.fetch_failed = msg;
-        }
-    }
+import { getSearchProvider } from "../providers/search-provider.js";
+async function handleSearch(input, engines, router, backendStatus, samplingServer, onProgress) {
+  const provider = await getSearchProvider();
+  return provider.search(input, {
+    engines,
+    router,
+    backendStatus,
+    samplingServer,
+    onProgress
+  });
 }
+export {
+  handleSearch
+};
 //# sourceMappingURL=search.js.map

package/dist/tools/search.js.map

@@ -1 +1 @@
-{"version":3,"file":"search.js","sourceRoot":"","sources":["../../src/tools/search.ts"],"names":[],"mappings":"AAEA,OAAO,EAAE,kBAAkB,EAA2B,MAAM,oBAAoB,CAAC;AACjF,OAAO,EAAE,cAAc,EAAE,MAAM,oBAAoB,CAAC;AACpD,OAAO,EAAE,aAAa,EAAE,MAAM,wBAAwB,CAAC;AACvD,OAAO,EAAE,aAAa,EAAE,MAAM,qBAAqB,CAAC;AACpD,OAAO,EAAE,eAAe,EAAE,MAAM,sBAAsB,CAAC;AACvD,OAAO,EAAE,cAAc,EAAE,MAAM,2BAA2B,CAAC;AAC3D,OAAO,EAAE,kBAAkB,EAAE,sBAAsB,EAAE,MAAM,mBAAmB,CAAC;AAC/E,OAAO,EAAE,SAAS,EAAE,MAAM,cAAc,CAAC;AACzC,OAAO,EAAE,YAAY,EAAE,MAAM,cAAc,CAAC;AAE5C,MAAM,GAAG,GAAG,YAAY,CAAC,QAAQ,CAAC,CAAC;AAEnC,MAAM,mBAAmB,GAAG,CAAC,CAAC;AAC9B,MAAM,eAAe,GAAG,EAAE,CAAC;AAC3B,MAAM,yBAAyB,GAAG,KAAK,CAAC;AACxC,MAAM,uBAAuB,GAAG,KAAK,CAAC;AAEtC,MAAM,CAAC,KAAK,UAAU,YAAY,CAChC,KAAkB,EAClB,OAAuB,EACvB,MAAmB;IAEnB,MAAM,KAAK,GAAG,IAAI,CAAC,GAAG,EAAE,CAAC;IACzB,MAAM,MAAM,GAAG,SAAS,EAAE,CAAC;IAE3B,MAAM,UAAU,GAAG,IAAI,CAAC,GAAG,CAAC,KAAK,CAAC,WAAW,IAAI,mBAAmB,EAAE,eAAe,CAAC,CAAC;IACvF,MAAM,cAAc,GAAG,KAAK,CAAC,eAAe,IAAI,IAAI,CAAC;IACrD,MAAM,eAAe,GAAG,KAAK,CAAC,iBAAiB,IAAI,yBAAyB,CAAC;IAC7E,MAAM,aAAa,GAAG,KAAK,CAAC,eAAe,IAAI,uBAAuB,CAAC;IACvE,MAAM,cAAc,GAAG,MAAM,CAAC,oBAAoB,CAAC;IACnD,MAAM,cAAc,GAAG,MAAM,CAAC,oBAAoB,CAAC;IAEnD,MAAM,MAAM,GAAG,sBAAsB,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;IACnD,IAAI,MAAM,IAAI,CAAC,cAAc,EAAE,CAAC;QAC9B,GAAG,CAAC,IAAI,CAAC,mCAAmC,EAAE,EAAE,KAAK,EAAE,KAAK,CAAC,KAAK,EAAE,CAAC,CAAC;QACtE,OAAO;YACL,OAAO,EAAE,MAAM,CAAC,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,UAAU,CAAC;YAC5C,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,YAAY,EAAE,MAAM,CAAC,YAAY;YACjC,aAAa,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;SAClC,CAAC;IACJ,CAAC;IAED,IAAI,aAAa,GAAG,OAAO,CAAC;IAC5B,IAAI,KAAK,CAAC,cAAc,IAAI,KAAK,CAAC,cAAc,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QAC5D,aAAa,GAAG,OAAO,CAAC,MAAM,CAAC,CAAC,CAAC,EAAE,CAAC,KAAK,CAAC,cAAe,CAAC,QAAQ,CAAC,CAAC,CAAC,IAAI,CAAC,CAAC,CAAC;QAC5E,IAAI,aAAa,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;YAC/B,GAAG,CAAC,IAAI,CAAC,qDAAqD,EAAE,EAAE,SAAS,EAAE,KAAK,CAAC,cAAc,EAAE,CAAC,CAAC;YACrG,aAAa,GAAG,OAAO,CAAC;QAC1B,CAAC;IACH,CAAC;IAED,MAAM,UAAU,GAAG,cAAc,CAAC,KAAK,CAAC,KAAK,CAAC,CAAC;IAC/C,GAAG,CAAC,KAAK,CAAC,qBAAqB,EAAE,EAAE,QAAQ,EAAE,KAAK,CAAC,KAAK,EAAE,KAAK,EAAE,UAAU,CAAC,MAAM,EAAE,CAAC,CAAC;IAEtF,MAAM,MAAM,GAAsB,EAAE,CAAC;IACrC,MAAM,WAAW,GAAG,IAAI,GAAG,EAAU,CAAC;IACtC,MAAM,MAAM,GAAa,EAAE,CAAC;IAE5B,wFAAwF;IACxF,MAAM,kBAAkB,GAAG,CAAC,CAAC,CAAC,KAAK,CAAC,eAAe,EAAE,MAAM,IAAI,KAAK,CAAC,eAAe,EAAE,MAAM,CAAC,CAAC;IAC9F,MAAM,eAAe,GAAG,kBAAkB,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC,CAAC;IAEnD,MAAM,cAAc,GAAG,aAAa,CAAC,OAAO,CAAC,MAAM,CAAC,EAAE,CACpD,UAAU,CAAC,GAAG,CAAC,KAAK,EAAE,KAAK,EAAE,EAAE;QAC7B,IAAI,CAAC;YACH,MAAM,OAAO,GAAG,MAAM,MAAM,CAAC,MAAM,CAAC,KAAK,EAAE;gBACzC,UAAU,EAAE,UAAU,GAAG,eAAe;gBACxC,SAAS,EAAE,KAAK,CAAC,UAAU;gBAC3B,QAAQ,EAAE,KAAK,CAAC,QAAQ;gBACxB,cAAc,EAAE,KAAK,CAAC,eAAe;gBACrC,cAAc,EAAE,KAAK,CAAC,eAAe;gBACrC,QAAQ,EAAE,KAAK,CAAC,SAAS;gBACzB,MAAM,EAAE,KAAK,CAAC,OAAO;gBACrB,QAAQ,EAAE,KAAK,CAAC,QAAQ;aACzB,CAAC,CAAC;YACH,KAAK,MAAM,CAAC,IAAI,OAAO,EAAE,CAAC;gBACxB,MAAM,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC;gBACf,WAAW,CAAC,GAAG,CAAC,MAAM,CAAC,IAAI,CAAC,CAAC;YAC/B,CAAC;QACH,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,MAAM,GAAG,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;YAC7D,GAAG,CAAC,IAAI,CAAC,sBAAsB,EAAE,EAAE,MAAM,EAAE,MAAM,CAAC,IAAI,EAAE,KAAK,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC;YAC7E,MAAM,CAAC,IAAI,CAAC,GAAG,MAAM,CAAC,IAAI,KAAK,GAAG,EAAE,CAAC,CAAC;QACxC,CAAC;IACH,CAAC,CAAC,CACH,CAAC;IAEF,MAAM,OAAO,CAAC,UAAU,CAAC,cAAc,CAAC,CAAC;IAEzC,IAAI,MAAM,CAAC,MAAM,KAAK,CAAC,EAAE,CAAC;QACxB,OAAO;YACL,OAAO,EAAE,EAAE;YACX,KAAK,EAAE,KAAK,CAAC,KAAK;YAClB,YAAY,EAAE,CAAC,GAAG,WAAW,CAAC;YAC9B,aAAa,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;YACjC,KAAK,EAAE,MAAM,CAAC,MAAM,GAAG,CAAC,CAAC,CAAC,CAAC,MAAM,CAAC,IAAI,CAAC,IAAI,CAAC,CAAC,CAAC,CAAC,kBAAkB;SAClE,CAAC;IACJ,CAAC;IAED,IAAI,MAAM,GAAG,kBAAkB,CAAC,MAAM,CAAC,CAAC;IAExC,0EAA0E;IAC1E,4EAA4E;IAC5E,MAAM,GAAG,eAAe,CAAC,MAAM,EAAE;QAC/B,cAAc,EAAE,KAAK,CAAC,eAAe;QACrC,cAAc,EAAE,KAAK,CAAC,eAAe;QACrC,QAAQ,EAAE,KAAK,CAAC,SAAS;QACzB,MAAM,EAAE,KAAK,CAAC,OAAO;QACrB,QAAQ,EAAE,KAAK,CAAC,QAAQ;KACzB,CAAC,CAAC;IAEH,MAAM,GAAG,MAAM,aAAa,CAAC,KAAK,CAAC,KAAK,EAAE,MAAM,CAAC,CAAC;IAClD,MAAM,GAAG,MAAM,aAAa,CAAC,MAAM,CAAC,CAAC;IAErC,MAAM,GAAG,MAAM,CAAC,KAAK,CAAC,CAAC,EAAE,UAAU,CAAC,CAAC;IAErC,MAAM,OAAO,GAAuB,MAAM,CAAC,GAAG,CAAC,CAAC,CAAC,EAAE,CAAC,CAAC;QACnD,KAAK,EAAE,CAAC,CAAC,KAAK;QACd,GAAG,EAAE,CAAC,CAAC,GAAG;QACV,OAAO,EAAE,CAAC,CAAC,OAAO;QAClB,eAAe,EAAE,CAAC,CAAC,eAAe;KACnC,CAAC,CAAC,CAAC;IAEJ,IAAI,cAAc,IAAI,OAAO,CAAC,MAAM,GAAG,CAAC,EAAE,CAAC;QACzC,MAAM,sBAAsB,CAAC,OAAO,EAAE,MAAM,EAAE;YAC5C,eAAe;YACf,aAAa;YACb,cAAc;YACd,aAAa,EAAE,KAAK,GAAG,cAAc;SACtC,CAAC,CAAC;IACL,CAAC;IAED,IAAI,CAAC;QACH,kBAAkB,CAAC,KAAK,CAAC,KAAK,EAAE,OAAO,EAAE,CAAC,GAAG,WAAW,CAAC,CAAC,CAAC;IAC7D,CAAC;IAAC,OAAO,GAAG,EAAE,CAAC;QACb,GAAG,CAAC,IAAI,CAAC,gCAAgC,EAAE,EAAE,KAAK,EAAE,MAAM,CAAC,GAAG,CAAC,EAAE,CAAC,CAAC;IACrE,CAAC;IAED,OAAO;QACL,OAAO;QACP,KAAK,EAAE,KAAK,CAAC,KAAK;QAClB,YAAY,EAAE,CAAC,GAAG,WAAW,CAAC;QAC9B,aAAa,EAAE,IAAI,CAAC,GAAG,EAAE,GAAG,KAAK;KAClC,CAAC;AACJ,CAAC;AASD,gHAAgH;AAChH,KAAK,UAAU,sBAAsB,CACnC,OAA2B,EAC3B,MAAmB,EACnB,GAAiB;IAEjB,IAAI,cAAc,GAAG,CAAC,CAAC;IAEvB,KAAK,MAAM,MAAM,IAAI,OAAO,EAAE,CAAC;QAC7B,IAAI,IAAI,CAAC,GAAG,EAAE,IAAI,GAAG,CAAC,aAAa,EAAE,CAAC;YACpC,MAAM,CAAC,YAAY,GAAG,eAAe,CAAC;YACtC,SAAS;QACX,CAAC;QAED,IAAI,cAAc,IAAI,GAAG,CAAC,aAAa,EAAE,CAAC;YACxC,MAAM,CAAC,iBAAiB,GAAG,IAAI,CAAC;YAChC,SAAS;QACX,CAAC;QAED,IAAI,CAAC;YACH,MAAM,GAAG,GAAG,MAAM,OAAO,CAAC,IAAI,CAAC;gBAC7B,MAAM,CAAC,KAAK,CAAC,MAAM,CAAC,GAAG,EAAE,EAAE,QAAQ,EAAE,MAAM,EAAE,CAAC;gBAC9C,IAAI,OAAO,CAAQ,CAAC,CAAC,EAAE,MAAM,EAAE,EAAE,CAC/B,UAAU,CAAC,GAAG,EAAE,CAAC,MAAM,CAAC,IAAI,KAAK,CAAC,SAAS,CAAC,CAAC,EAAE,GAAG,CAAC,cAAc,CAAC,CACnE;aACF,CAAC,CAAC;YAEH,MAAM,UAAU,GAAG,MAAM,cAAc,CAAC,GAAG,CAAC,IAAI,EAAE,GAAG,CAAC,QAAQ,EAAE;gBAC9D,QAAQ,EAAE,GAAG,CAAC,eAAe;gBAC7B,WAAW,EAAE,GAAG,CAAC,WAAW;aAC7B,CAAC,CAAC;YAEH,IAAI,OAAO,GAAG,UAAU,CAAC,QAAQ,CAAC;YAElC,MAAM,SAAS,GAAG,GAAG,CAAC,aAAa,GAAG,cAAc,CAAC;YACrD,IAAI,OAAO,CAAC,MAAM,GAAG,SAAS,EAAE,CAAC;gBAC/B,OAAO,GAAG,OAAO,CAAC,KAAK,CAAC,CAAC,EAAE,SAAS,CAAC,CAAC;gBACtC,MAAM,CAAC,iBAAiB,GAAG,IAAI,CAAC;YAClC,CAAC;YAED,cAAc,IAAI,OAAO,CAAC,MAAM,CAAC;YACjC,MAAM,CAAC,gBAAgB,GAAG,OAAO,CAAC;QACpC,CAAC;QAAC,OAAO,GAAG,EAAE,CAAC;YACb,MAAM,GAAG,GAAG,GAAG,YAAY,KAAK,CAAC,CAAC,CAAC,GAAG,CAAC,OAAO,CAAC,CAAC,CAAC,MAAM,CAAC,GAAG,CAAC,CAAC;YAC7D,GAAG,CAAC,KAAK,CAAC,sBAAsB,EAAE,EAAE,GAAG,EAAE,MAAM,CAAC,GAAG,EAAE,KAAK,EAAE,GAAG,EAAE,CAAC,CAAC;YACnE,MAAM,CAAC,YAAY,GAAG,GAAG,CAAC;QAC5B,CAAC;IACH,CAAC;AACH,CAAC"}
+{"version":3,"sources":["../../src/tools/search.ts"],"sourcesContent":["import type {\n  SearchInput,\n  SearchOutput,\n  SearchEngine,\n  ProgressCallback,\n  StageResult,\n} from '../types.js';\nimport type { SmartRouter } from '../fetch/router.js';\nimport type { BackendStatus } from '../server/backend-status.js';\nimport type { SamplingCapableServer } from '../search/sampling.js';\nimport { getSearchProvider } from '../providers/search-provider.js';\n\n/**\n * Thin handler — validates wiring args and delegates to the selected search\n * provider (legacy SearXNG by default, V1 stub when `WIGOLO_SEARCH=v1`).\n *\n * All orchestration logic lives in `src/search/legacy/searxng-orchestrator.ts`.\n */\nexport async function handleSearch(\n  input: SearchInput,\n  engines: SearchEngine[],\n  router: SmartRouter,\n  backendStatus?: BackendStatus,\n  samplingServer?: SamplingCapableServer,\n  onProgress?: ProgressCallback,\n): Promise<StageResult<SearchOutput>> {\n  const provider = await getSearchProvider();\n  return provider.search(input, {\n    engines,\n    router,\n    backendStatus,\n    samplingServer,\n    onProgress,\n  });\n}\n"],"mappings":"AAUA,SAAS,yBAAyB;AAQlC,eAAsB,aACpB,OACA,SACA,QACA,eACA,gBACA,YACoC;AACpC,QAAM,WAAW,MAAM,kBAAkB;AACzC,SAAO,SAAS,OAAO,OAAO;AAAA,IAC5B;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,IACA;AAAA,EACF,CAAC;AACH;","names":[]}

package/dist/types.d.ts

@@ -1,13 +1,64 @@
 import type { JsonSchema } from './extraction/schema.js';
+export type Mode = 'cache' | 'default' | 'stealth';
+export declare const MODES: readonly Mode[];
+export type DeprecatedMode = 'fast' | 'balanced' | 'deep';
+export declare const DEPRECATED_MODES: readonly DeprecatedMode[];
+export interface StageError {
+    error: string;
+    error_reason: string;
+    stage: string;
+    hint?: string;
+}
+export type StageResult<T> = {
+    ok: true;
+    data: T;
+} | ({
+    ok: false;
+} & StageError);
+export type BrowserAction = {
+    type: 'click';
+    selector: string;
+} | {
+    type: 'type';
+    selector: string;
+    text: string;
+} | {
+    type: 'wait';
+    ms: number;
+} | {
+    type: 'wait_for';
+    selector: string;
+    timeout?: number;
+} | {
+    type: 'scroll';
+    direction: 'down' | 'up';
+    amount?: number;
+} | {
+    type: 'screenshot';
+};
+export interface ActionResult {
+    action_index: number;
+    type: BrowserAction['type'];
+    success: boolean;
+    error?: string;
+    screenshot?: string;
+}
 export interface FetchInput {
     url: string;
     render_js?: 'auto' | 'always' | 'never';
     use_auth?: boolean;
     max_chars?: number;
+    max_content_chars?: number;
     section?: string;
     section_index?: number;
     screenshot?: boolean;
     headers?: Record<string, string>;
+    actions?: BrowserAction[];
+    force_refresh?: boolean;
+    max_tokens_out?: number;
+    include_full_markdown?: boolean;
+    citation_format?: CitationFormat;
+    mode?: Mode;
 }
 export interface FetchOutput {
     url: string;
@@ -18,13 +69,25 @@ export interface FetchOutput {
         author?: string;
         date?: string;
         language?: string;
+        og_image?: string;
+        og_type?: string;
+        canonical_url?: string;
+        keywords?: string[];
         section_matched?: boolean;
     };
     links: string[];
     images: string[];
     screenshot?: string;
     cached: boolean;
+    cached_at?: string;
+    stale?: boolean;
+    js_required?: boolean;
     error?: string;
+    action_results?: ActionResult[];
+    changed?: boolean;
+    previous_hash?: string;
+    diff_summary?: string;
+    evidence?: EvidenceItem[];
 }
 export interface RawFetchResult {
     url: string;
@@ -36,6 +99,9 @@ export interface RawFetchResult {
     headers: Record<string, string>;
     rawBuffer?: Buffer;
     screenshot?: string;
+    actionResults?: ActionResult[];
+    jsRequired?: boolean;
+    escalated?: boolean;
 }
 export interface ExtractionResult {
     title: string;
@@ -45,18 +111,24 @@ export interface ExtractionResult {
         author?: string;
         date?: string;
         language?: string;
+        og_image?: string;
+        og_type?: string;
+        canonical_url?: string;
+        keywords?: string[];
     };
     links: string[];
     images: string[];
     extractor: ExtractorType;
 }
-export type ExtractorType = 'defuddle' | 'readability' | 'turndown' | 'site-specific' | 'trafilatura';
-export type BrowserType = 'chromium' | 'firefox' | 'webkit';
+export type ExtractorType = 'defuddle' | 'readability' | 'turndown' | 'site-specific';
+export type BrowserType = 'chromium' | 'firefox' | 'webkit' | 'lightpanda';
 export interface CDPSession {
     id: string;
     url: string;
     title: string;
     webSocketDebuggerUrl: string;
+    type?: string;
+    devtoolsFrontendUrl?: string;
 }
 export interface CachedContent {
     id: number;
@@ -79,12 +151,36 @@ export interface Extractor {
     canHandle(url: string, html?: string): boolean;
     extract(html: string, url: string): ExtractionResult | null;
 }
+export type FieldProvenance = 'json-ld' | 'microdata' | 'rdfa' | 'heuristic' | 'llm';
+export interface StructuredDataResult {
+    provenance: 'json-ld' | 'microdata' | 'rdfa';
+    type: string;
+    fields: Record<string, unknown>;
+}
+export interface SchemaExtractionResult {
+    values: Record<string, unknown>;
+    provenance: Record<string, FieldProvenance>;
+}
+/**
+ * Optional agent-supplied context that the v1 search pipeline uses to refine
+ * ranking and dedup. All fields are advisory — the engine never requires them.
+ */
+export interface AgentContext {
+    /** Surrounding code, prior assistant turn, or task framing. */
+    text?: string;
+    /** URLs the agent has already seen recently; v1 dedups against these. */
+    recent_urls?: string[];
+    /** Optional one-line task framing — included in the embedded query if `text` empty. */
+    intent?: string;
+}
 export interface SearchInput {
-    query: string;
+    query: string | string[];
     max_results?: number;
     include_content?: boolean;
     content_max_chars?: number;
+    max_content_chars?: number;
     max_total_chars?: number;
+    max_fetches?: number;
     time_range?: 'day' | 'week' | 'month' | 'year';
     search_engines?: string[];
     language?: string;
@@ -93,6 +189,14 @@ export interface SearchInput {
     from_date?: string;
     to_date?: string;
     category?: 'general' | 'news' | 'code' | 'docs' | 'papers' | 'images';
+    format?: 'answer' | 'stream_answer';
+    max_highlights?: number;
+    force_refresh?: boolean;
+    max_tokens_out?: number;
+    include_full_markdown?: boolean;
+    citation_format?: CitationFormat;
+    mode?: Mode;
+    agent_context?: AgentContext;
 }
 export interface SearchResultItem {
     title: string;
@@ -102,13 +206,159 @@ export interface SearchResultItem {
     fetch_failed?: string;
     content_truncated?: boolean;
     relevance_score: number;
+    published_date?: string;
+    cached?: boolean;
+    cached_at?: string;
+    stale?: boolean;
 }
 export interface SearchOutput {
     results: SearchResultItem[];
     query: string;
     engines_used: string[];
     total_time_ms: number;
+    search_time_ms?: number;
+    fetch_time_ms?: number;
+    error?: string;
+    warning?: string;
+    context_text?: string;
+    queries_executed?: string[];
+    answer?: string;
+    citations?: Citation[];
+    highlights?: Highlight[];
+    streaming?: boolean;
+    evidence?: EvidenceItem[];
+    citations_xml?: string;
+}
+export interface SourceSpan {
+    start: number;
+    end: number;
+}
+export interface EvidenceItem {
+    title: string;
+    url: string;
+    section_heading: string | null;
+    excerpt: string;
+    score: number;
+    citation_id: string;
+    source_span: SourceSpan;
+}
+export type CitationFormat = 'numbered' | 'anthropic_tags' | 'json';
+export interface Highlight {
+    text: string;
+    source_index: number;
+    relevance_score: number;
+    source_url: string;
+    source_title: string;
+    section_heading?: string | null;
+    source_span?: SourceSpan;
+}
+export interface Citation {
+    index: number;
+    url: string;
+    title: string;
+    snippet: string;
+    citation_id?: string;
+}
+export interface ProgressUpdate {
+    progress: number;
+    total?: number;
+    message?: string;
+}
+export type ProgressCallback = (update: ProgressUpdate) => void | Promise<void>;
+export interface ResearchInput {
+    question: string;
+    depth?: 'quick' | 'standard' | 'comprehensive';
+    max_sources?: number;
+    include_domains?: string[];
+    exclude_domains?: string[];
+    schema?: Record<string, unknown>;
+    stream?: boolean;
+    max_tokens_out?: number;
+    include_full_markdown?: boolean;
+    citation_format?: CitationFormat;
+}
+export interface ResearchSource {
+    url: string;
+    title: string;
+    markdown_content: string;
+    relevance_score: number;
+    fetched: boolean;
+    fetch_error?: string;
+}
+export interface ResearchOutput {
+    report: string;
+    citations: Citation[];
+    sources: ResearchSource[];
+    sub_queries: string[];
+    depth: string;
+    total_time_ms: number;
+    sampling_supported: boolean;
+    brief?: ResearchBrief;
+    error?: string;
+    evidence?: EvidenceItem[];
+}
+export interface CrossReference {
+    finding: string;
+    source_indices: number[];
+    confidence: 'high' | 'medium';
+}
+export interface CitationGraphEntry {
+    claim: string;
+    source_indices: number[];
+    confidence: 'high' | 'medium' | 'low';
+}
+export interface ResearchBrief {
+    topics: string[];
+    highlights: Highlight[];
+    key_findings: string[];
+    per_source_char_cap: number;
+    total_sources_char_cap: number;
+    sections: {
+        overview: {
+            key_findings: string[];
+            cross_references: CrossReference[];
+        };
+        comparison?: {
+            entities: string[];
+            comparison_points: string[];
+        };
+        gaps: string[];
+    };
+    query_type: 'comparison' | 'how-to' | 'concept' | 'general';
+    citation_graph?: CitationGraphEntry[];
+}
+export interface AgentInput {
+    prompt: string;
+    urls?: string[];
+    schema?: Record<string, unknown>;
+    max_pages?: number;
+    max_time_ms?: number;
+    stream?: boolean;
+    max_tokens_out?: number;
+    include_full_markdown?: boolean;
+    citation_format?: CitationFormat;
+}
+export interface AgentSource {
+    url: string;
+    title: string;
+    markdown_content: string;
+    fetched: boolean;
+    fetch_error?: string;
+}
+export interface AgentStep {
+    action: 'plan' | 'search' | 'fetch' | 'extract' | 'synthesize';
+    detail: string;
+    time_ms: number;
+}
+export interface AgentOutput {
+    result: string | Record<string, unknown>;
+    sources: AgentSource[];
+    pages_fetched: number;
+    steps: AgentStep[];
+    total_time_ms: number;
+    sampling_supported: boolean;
     error?: string;
+    evidence?: EvidenceItem[];
 }
 export interface RawSearchResult {
     title: string;
@@ -116,6 +366,7 @@ export interface RawSearchResult {
     snippet: string;
     relevance_score: number;
     engine: string;
+    published_date?: string;
 }
 export interface SearchEngineOptions {
     maxResults?: number;
@@ -136,18 +387,23 @@ export interface CrawlInput {
     url: string;
     max_depth?: number;
     max_pages?: number;
-    strategy?: 'bfs' | 'dfs' | 'sitemap' | 'map';
+    strategy?: 'bfs' | 'dfs' | 'sitemap' | 'map' | 'auto';
     include_patterns?: string[];
     exclude_patterns?: string[];
     use_auth?: boolean;
     extract_links?: boolean;
     max_total_chars?: number;
+    max_tokens_out?: number;
+    include_full_markdown?: boolean;
+    citation_format?: CitationFormat;
 }
 export interface CrawlResultItem {
     url: string;
     title: string;
     markdown: string;
     depth: number;
+    evidence?: EvidenceItem[];
+    excerpt?: string;
 }
 export interface LinkEdge {
     from: string;
@@ -157,6 +413,8 @@ export interface CrawlOutput {
     pages: CrawlResultItem[];
     total_found: number;
     crawled: number;
+    /** Pages fetched but excluded from `pages` due to max_total_chars budget. */
+    dropped_over_budget?: number;
     links?: LinkEdge[];
     error?: string;
 }
@@ -172,6 +430,16 @@ export interface CacheInput {
     since?: string;
     clear?: boolean;
     stats?: boolean;
+    check_changes?: boolean;
+    /**
+     * Search strategy when `query` is provided:
+     *   - 'fts'    (default) keyword-only BM25 over FTS5
+     *   - 'hybrid' BM25 + semantic vector search fused with reciprocal rank fusion
+     * Hybrid mode requires the sqlite-vec extension and a populated embedding
+     * index; on miss it transparently falls back to 'fts'.
+     */
+    mode?: 'fts' | 'hybrid';
+    limit?: number;
 }
 export interface CacheResultItem {
     url: string;
@@ -190,14 +458,26 @@ export interface CacheOutput {
     stats?: CacheStats;
     cleared?: number;
     error?: string;
+    changes?: ChangeReport[];
+}
+export interface ChangeReport {
+    url: string;
+    changed: boolean;
+    previous_hash?: string;
+    current_hash?: string;
+    diff_summary?: string;
+    error?: string;
 }
+export type NamedSchemaType = 'Article' | 'Recipe' | 'Product' | 'CodeSnippet' | 'Paper' | 'EventListing';
 export interface ExtractInput {
     url?: string;
     html?: string;
-    mode?: 'selector' | 'tables' | 'metadata' | 'schema';
+    mode?: 'selector' | 'tables' | 'metadata' | 'schema' | 'structured';
     css_selector?: string;
     multiple?: boolean;
     schema?: JsonSchema;
+    named_schema?: NamedSchemaType;
+    execution_mode?: Mode;
 }
 export interface MetadataData {
     title?: string;
@@ -206,6 +486,8 @@ export interface MetadataData {
     date?: string;
     keywords?: string[];
     og_image?: string;
+    og_type?: string;
+    canonical_url?: string;
     jsonld?: Record<string, unknown>[];
 }
 export interface TableData {
@@ -213,10 +495,71 @@ export interface TableData {
     headers: string[];
     rows: Array<Record<string, string>>;
 }
+export interface DefinitionPair {
+    term: string;
+    description: string;
+}
+export interface ChartHint {
+    title?: string;
+    aria_label?: string;
+    figcaption?: string;
+    type_hint?: 'chart' | 'diagram' | 'graph';
+}
+export interface KeyValuePair {
+    key: string;
+    value: string;
+    source: 'microdata' | 'data-attr' | 'comparison-grid' | 'text-pattern';
+}
+export interface StructuredData {
+    tables: TableData[];
+    definitions: DefinitionPair[];
+    jsonld: Record<string, unknown>[];
+    chart_hints: ChartHint[];
+    key_value_pairs: KeyValuePair[];
+}
 export interface ExtractOutput {
-    data: string | string[] | TableData[] | MetadataData | Record<string, unknown>;
+    data: string | string[] | TableData[] | MetadataData | StructuredData | Record<string, unknown>;
     source_url?: string;
-    mode: 'selector' | 'tables' | 'metadata' | 'schema';
+    mode: 'selector' | 'tables' | 'metadata' | 'schema' | 'structured';
+    error?: string;
+    warnings?: string[];
+}
+export interface MatchSignals {
+    embedding_rank?: number;
+    fts5_rank?: number;
+    fused_score: number;
+}
+export interface FindSimilarResult {
+    url: string;
+    title: string;
+    markdown: string;
+    relevance_score: number;
+    source: 'cache' | 'search';
+    match_signals: MatchSignals;
+}
+export interface FindSimilarInput {
+    url?: string;
+    concept?: string;
+    max_results?: number;
+    include_domains?: string[];
+    exclude_domains?: string[];
+    include_cache?: boolean;
+    include_web?: boolean;
+    mode?: 'auto' | 'cache' | 'web-expansion' | 'crawl-rank';
+    max_tokens_out?: number;
+    include_full_markdown?: boolean;
+    citation_format?: CitationFormat;
+}
+export interface FindSimilarOutput {
+    results: FindSimilarResult[];
+    method: 'hybrid' | 'embedding' | 'fts5' | 'search';
+    cache_hits: number;
+    search_hits: number;
+    embedding_available: boolean;
+    cold_start?: string;
+    cache_seeded?: boolean;
     error?: string;
+    total_time_ms: number;
+    evidence?: EvidenceItem[];
 }
 //# sourceMappingURL=types.d.ts.map