@staticn0va/wigolo 0.1.0 → 0.1.2

package/dist/server.js

@@ -1,311 +1,468 @@
-import { join } from 'node:path';
-import { mkdirSync } from 'node:fs';
-import { Server } from '@modelcontextprotocol/sdk/server/index.js';
-import { StdioServerTransport } from '@modelcontextprotocol/sdk/server/stdio.js';
-import { ListToolsRequestSchema, CallToolRequestSchema, } from '@modelcontextprotocol/sdk/types.js';
-import { SmartRouter } from './fetch/router.js';
-import { BrowserPool } from './fetch/browser-pool.js';
-import { httpFetch } from './fetch/http-client.js';
-import { initDatabase, closeDatabase } from './cache/db.js';
-import { handleFetch } from './tools/fetch.js';
-import { handleSearch } from './tools/search.js';
-import { handleCrawl } from './tools/crawl.js';
-import { handleCache } from './tools/cache.js';
-import { handleExtract } from './tools/extract.js';
-import { SearxngClient } from './search/searxng.js';
-import { DuckDuckGoEngine } from './search/engines/duckduckgo.js';
-import { BingEngine } from './search/engines/bing.js';
-import { StartpageEngine } from './search/engines/startpage.js';
-import { resolveSearchBackend } from './searxng/bootstrap.js';
-import { SearxngProcess } from './searxng/process.js';
-import { DockerSearxng } from './searxng/docker.js';
-import { getConfig } from './config.js';
-import { createLogger } from './logger.js';
-const log = createLogger('server');
-const FETCH_TOOL_SCHEMA = {
-    type: 'object',
-    properties: {
-        url: { type: 'string', description: 'URL to fetch' },
-        render_js: {
-            type: 'string',
-            enum: ['auto', 'always', 'never'],
-            description: 'JavaScript rendering mode (default: auto)',
-        },
-        use_auth: {
-            type: 'boolean',
-            description: 'Use stored auth credentials (default: false)',
-        },
-        max_chars: {
-            type: 'number',
-            description: 'Maximum characters to return',
-        },
-        section: {
-            type: 'string',
-            description: 'Extract a specific section by heading text',
-        },
-        section_index: {
-            type: 'number',
-            description: 'Index of the section match (default: 0)',
-        },
-        screenshot: {
-            type: 'boolean',
-            description: 'Capture a screenshot (default: false)',
-        },
-        headers: {
-            type: 'object',
-            description: 'Additional HTTP headers',
-            additionalProperties: { type: 'string' },
-        },
-    },
-    required: ['url'],
-};
-const SEARCH_TOOL_SCHEMA = {
-    type: 'object',
-    properties: {
-        query: { type: 'string', description: 'Search query' },
-        max_results: { type: 'number', description: 'Max results to return (default 5, max 20)' },
-        include_content: { type: 'boolean', description: 'Fetch full content for results (default true)' },
-        content_max_chars: { type: 'number', description: 'Max chars per result content (default 30000)' },
-        max_total_chars: { type: 'number', description: 'Max total chars across all results (default 50000)' },
-        time_range: { type: 'string', enum: ['day', 'week', 'month', 'year'], description: 'Time range filter' },
-        search_engines: { type: 'array', items: { type: 'string' }, description: 'Override engine selection' },
-        language: { type: 'string', description: 'Language preference' },
-        include_domains: {
-            type: 'array',
-            items: { type: 'string' },
-            description: 'Only return results from these domains (e.g. ["react.dev", "github.com"])',
-        },
-        exclude_domains: {
-            type: 'array',
-            items: { type: 'string' },
-            description: 'Never return results from these domains',
-        },
-        from_date: {
-            type: 'string',
-            description: 'ISO date (YYYY-MM-DD) — only return results published after this date',
-        },
-        to_date: {
-            type: 'string',
-            description: 'ISO date (YYYY-MM-DD) — only return results published before this date',
-        },
-        category: {
-            type: 'string',
-            enum: ['general', 'news', 'code', 'docs', 'papers', 'images'],
-            description: 'Category of search (general, news, code, docs, papers, images)',
-        },
-    },
-    required: ['query'],
-};
-const CRAWL_TOOL_SCHEMA = {
-    type: 'object',
-    properties: {
-        url: { type: 'string', description: 'Seed URL to start crawling from' },
-        max_depth: { type: 'number', description: 'Maximum link depth from seed (default: 2)' },
-        max_pages: { type: 'number', description: 'Maximum pages to crawl (default: 20)' },
-        strategy: {
-            type: 'string',
-            enum: ['bfs', 'dfs', 'sitemap', 'map'],
-            description: 'Crawl strategy: bfs (breadth-first), dfs (depth-first), sitemap (use sitemap.xml), map (URL-only discovery — returns list of URLs without content, faster than full crawl)',
-        },
-        include_patterns: {
-            type: 'array',
-            items: { type: 'string' },
-            description: 'URL regex whitelist — only crawl matching URLs',
-        },
-        exclude_patterns: {
-            type: 'array',
-            items: { type: 'string' },
-            description: 'URL regex blacklist — skip matching URLs',
-        },
-        use_auth: { type: 'boolean', description: 'Use stored auth credentials (default: false)' },
-        extract_links: { type: 'boolean', description: 'Return link graph between pages (default: false)' },
-        max_total_chars: { type: 'number', description: 'Max total chars across all pages (default: 100000)' },
-    },
-    required: ['url'],
-};
-const CACHE_TOOL_SCHEMA = {
-    type: 'object',
-    properties: {
-        query: { type: 'string', description: 'Full-text search over cached content' },
-        url_pattern: {
-            type: 'string',
-            description: 'Filter by URL glob pattern (e.g., "*example.com*")',
-        },
-        since: {
-            type: 'string',
-            description: 'ISO date — only results cached after this date',
-        },
-        clear: {
-            type: 'boolean',
-            description: 'Clear matching cache entries (requires at least one filter: query, url_pattern, or since)',
-        },
-        stats: {
-            type: 'boolean',
-            description: 'Return cache statistics (total URLs, size, date range)',
-        },
-    },
-};
-const EXTRACT_TOOL_SCHEMA = {
-    type: 'object',
-    properties: {
-        url: { type: 'string', description: 'URL to fetch and extract from' },
-        html: { type: 'string', description: 'Raw HTML to extract from (url takes priority if both provided)' },
-        mode: {
-            type: 'string',
-            enum: ['selector', 'tables', 'metadata', 'schema'],
-            description: 'Extraction mode: selector (CSS), tables (HTML tables), metadata (meta tags + JSON-LD), schema (extract fields matching a JSON Schema via heuristic matching)',
-        },
-        css_selector: {
-            type: 'string',
-            description: 'CSS selector to match (required when mode="selector")',
-        },
-        multiple: {
-            type: 'boolean',
-            description: 'Return array of all matches instead of first (default: false, only for mode="selector")',
-        },
-        schema: {
-            type: 'object',
-            description: 'JSON Schema defining fields to extract. Field names are matched against page content via CSS classes, ARIA labels, microdata, and JSON-LD. Required when mode="schema".',
-        },
-    },
-};
-export async function startServer() {
-    const config = getConfig();
-    mkdirSync(config.dataDir, { recursive: true });
-    initDatabase(join(config.dataDir, 'wigolo.db'));
-    const httpClient = {
-        fetch: (url, options) => httpFetch(url, options),
-    };
-    const browserPool = new BrowserPool();
-    const router = new SmartRouter(httpClient, browserPool);
-    // --- Search backend initialization ---
-    const backend = await resolveSearchBackend();
-    const searchEngines = [];
-    let searxngProcess = null;
-    let dockerSearxng = null;
-    if (backend.type === 'external' && backend.url) {
-        searchEngines.push(new SearxngClient(backend.url));
+import { dirname, join } from "node:path";
+import { mkdirSync, readFileSync } from "node:fs";
+import { fileURLToPath } from "node:url";
+import { Server } from "@modelcontextprotocol/sdk/server/index.js";
+import { StdioServerTransport } from "@modelcontextprotocol/sdk/server/stdio.js";
+import {
+  ListToolsRequestSchema,
+  CallToolRequestSchema,
+  ListResourcesRequestSchema,
+  ReadResourceRequestSchema
+} from "@modelcontextprotocol/sdk/types.js";
+import { SmartRouter } from "./fetch/router.js";
+import { MultiBrowserPool } from "./fetch/browser-pool.js";
+import { closeDaemonBrowser } from "./fetch/playwright-tier.js";
+import { httpFetch } from "./fetch/http-client.js";
+import { initDatabase, closeDatabase } from "./cache/db.js";
+import { handleFetch } from "./tools/fetch.js";
+import { handleSearch } from "./tools/search.js";
+import { handleCrawl } from "./tools/crawl.js";
+import { handleCache } from "./tools/cache.js";
+import { handleExtract } from "./tools/extract.js";
+import { handleFindSimilar } from "./tools/find-similar.js";
+import { handleResearch } from "./tools/research.js";
+import { handleAgent } from "./tools/agent.js";
+import { SearxngClient } from "./search/searxng.js";
+import { DuckDuckGoEngine } from "./search/engines/duckduckgo.js";
+import { BingEngine } from "./search/engines/bing.js";
+import { StartpageEngine } from "./search/engines/startpage.js";
+import { resolveSearchBackend, bootstrapNativeSearxng, getBootstrapState } from "./searxng/bootstrap.js";
+import { SearxngProcess } from "./searxng/process.js";
+import { DockerSearxng } from "./searxng/docker.js";
+import { BackendStatus } from "./server/backend-status.js";
+import { maybeEagerWarmup } from "./server/warmup-on-start.js";
+import { getEmbeddingService, resetEmbeddingService } from "./embedding/embed.js";
+import { getConfig } from "./config.js";
+import { createLogger } from "./logger.js";
+import {
+  WIGOLO_INSTRUCTIONS,
+  WIGOLO_INSTRUCTIONS_FULL,
+  WIGOLO_DOCS_URI,
+  TOOL_DESCRIPTIONS
+} from "./instructions.js";
+import {
+  FETCH_TOOL_SCHEMA,
+  SEARCH_TOOL_SCHEMA,
+  CRAWL_TOOL_SCHEMA,
+  CACHE_TOOL_SCHEMA,
+  EXTRACT_TOOL_SCHEMA,
+  FIND_SIMILAR_TOOL_SCHEMA,
+  RESEARCH_TOOL_SCHEMA,
+  AGENT_TOOL_SCHEMA
+} from "./server/tool-schemas.js";
+import { loadPlugins } from "./plugins/loader.js";
+import { PluginRegistry } from "./plugins/registry.js";
+import { registerExtractor } from "./extraction/pipeline.js";
+const log = createLogger("server");
+function readPackageVersion() {
+  try {
+    const here = dirname(fileURLToPath(import.meta.url));
+    const pkgPath = join(here, "..", "package.json");
+    const pkg = JSON.parse(readFileSync(pkgPath, "utf-8"));
+    return pkg.version ?? "0.0.0";
+  } catch {
+    return "0.0.0";
+  }
+}
+const SERVER_VERSION = readPackageVersion();
+async function initSubsystems() {
+  const config = getConfig();
+  mkdirSync(config.dataDir, { recursive: true });
+  initDatabase(join(config.dataDir, "wigolo.db"));
+  try {
+    await getEmbeddingService().init();
+  } catch (err) {
+    log.warn("embedding service init failed, find_similar will run without embedding path", {
+      error: String(err)
+    });
+  }
+  const httpClient = {
+    fetch: (url, options) => httpFetch(url, options)
+  };
+  const browserPool = new MultiBrowserPool({
+    browserTypes: config.browserTypes,
+    selectionStrategy: "round-robin"
+  });
+  const router = new SmartRouter(httpClient, browserPool);
+  if (config.lightpandaEnabled && config.lightpandaUrl) {
+    log.info("lightpanda browser backend enabled", {
+      url: config.lightpandaUrl,
+      failureThreshold: config.lightpandaFailureThreshold
+    });
+  }
+  const backendStatus = new BackendStatus();
+  const searchEngines = [
+    new BingEngine(),
+    new DuckDuckGoEngine(),
+    new StartpageEngine()
+  ];
+  const pluginRegistry = new PluginRegistry();
+  try {
+    const pluginResult = await loadPlugins();
+    for (const ext of pluginResult.extractors) {
+      pluginRegistry.registerExtractor(ext, ext.name);
+      registerExtractor(ext);
     }
-    else if (backend.type === 'native' && backend.searxngPath) {
-        searxngProcess = new SearxngProcess(backend.searxngPath, config.dataDir);
-        const url = await searxngProcess.start();
-        if (url) {
-            searchEngines.push(new SearxngClient(url));
+    for (const eng of pluginResult.searchEngines) {
+      pluginRegistry.registerSearchEngine(eng, eng.name);
+      searchEngines.push(eng);
+    }
+    if (pluginResult.errors.length > 0) {
+      log.warn("some plugins failed to load", {
+        errors: pluginResult.errors.map((e) => `${e.pluginName}: ${e.message}`)
+      });
+    }
+    if (pluginResult.loaded.length > 0) {
+      log.info("plugins loaded", {
+        count: pluginResult.loaded.length,
+        names: pluginResult.loaded.map((p) => p.name)
+      });
+    }
+  } catch (err) {
+    log.error("plugin loading failed", { error: String(err) });
+  }
+  let searxngProcess = null;
+  let dockerSearxng = null;
+  let searxngBootstrap = null;
+  async function bootstrapSearxng() {
+    try {
+      const initialState = getBootstrapState(config.dataDir);
+      if (!config.searxngUrl && initialState?.status !== "ready") {
+        backendStatus.markBootstrapping();
+      }
+      const backend = await resolveSearchBackend();
+      if (backend.type === "external" && backend.url) {
+        searchEngines.unshift(new SearxngClient(backend.url));
+        backendStatus.markHealthy();
+        log.info("using external search engine", { url: backend.url });
+        return;
+      }
+      if (backend.type === "native" && backend.searxngPath) {
+        const state = getBootstrapState(config.dataDir);
+        if (state?.status !== "ready") {
+          log.info("search engine not ready \u2014 bootstrapping in background; search uses fallback engines until ready");
+          try {
+            await bootstrapNativeSearxng(config.dataDir);
+          } catch (err) {
+            log.warn("search engine bootstrap failed, continuing with fallback scraping");
+            backendStatus.markUnhealthy(`bootstrap exception: ${String(err)}`);
+            return;
+          }
         }
-        else {
-            log.warn('SearXNG failed to start, using direct scraping fallback');
+        const postBootstrapState = getBootstrapState(config.dataDir);
+        if (postBootstrapState?.status === "ready") {
+          searxngProcess = new SearxngProcess(backend.searxngPath, config.dataDir, {
+            onUnhealthy: (reason) => {
+              backendStatus.markUnhealthy(reason);
+              const idx = searchEngines.findIndex((e) => e.name === "searxng");
+              if (idx >= 0) searchEngines.splice(idx, 1);
+              log.warn("search engine marked unhealthy", { reason });
+            },
+            onHealthy: () => {
+              const url2 = searxngProcess?.getUrl();
+              if (!url2) return;
+              backendStatus.markHealthy();
+              if (!searchEngines.some((e) => e.name === "searxng")) {
+                searchEngines.unshift(new SearxngClient(url2));
+              }
+              log.info("search engine recovered");
+            }
+          });
+          const url = await searxngProcess.start();
+          if (url) {
+            searchEngines.unshift(new SearxngClient(url));
+            backendStatus.markHealthy();
+            log.info("search engine ready", { url });
+          } else {
+            log.warn("search engine failed to start, using fallback scraping");
+            backendStatus.markUnhealthy("search engine process failed to start");
+          }
         }
-    }
-    else if (backend.type === 'docker') {
+        return;
+      }
+      if (backend.type === "docker") {
         dockerSearxng = new DockerSearxng();
         const url = await dockerSearxng.start();
         if (url) {
-            searchEngines.push(new SearxngClient(url));
-        }
-        else {
-            log.warn('Docker SearXNG failed to start, using direct scraping fallback');
+          searchEngines.unshift(new SearxngClient(url));
+          backendStatus.markHealthy();
+          log.info("search engine (docker) ready", { url });
+        } else {
+          log.warn("search engine (docker) failed to start, using fallback scraping");
+          backendStatus.markUnhealthy("search engine (docker) failed to start");
         }
+      }
+      if (backend.type === "scraping") {
+        const state = getBootstrapState(config.dataDir);
+        const reason = state?.lastError?.message ?? state?.error ?? "no search engine backend available";
+        backendStatus.markUnhealthy(reason);
+      }
+    } catch (err) {
+      log.warn("background backend setup failed", { error: String(err) });
+      backendStatus.markUnhealthy(`backend setup failed: ${String(err)}`);
     }
-    searchEngines.push(new BingEngine(), new DuckDuckGoEngine(), new StartpageEngine());
-    const server = new Server({ name: 'wigolo', version: '0.1.0' }, { capabilities: { tools: {} } });
-    server.setRequestHandler(ListToolsRequestSchema, async () => ({
-        tools: [
-            {
-                name: 'fetch',
-                description: 'Fetch a web page and return its content as clean markdown. ' +
-                    'Supports JavaScript rendering, auth, section extraction, and caching.',
-                inputSchema: FETCH_TOOL_SCHEMA,
-            },
-            {
-                name: 'search',
-                description: 'Search the web and return results with optional full content extraction. ' +
-                    'One call: query in, clean markdown out.',
-                inputSchema: SEARCH_TOOL_SCHEMA,
-            },
-            {
-                name: 'crawl',
-                description: 'Crawl a website starting from a seed URL. Supports BFS, DFS, and sitemap strategies ' +
-                    'with depth/page limits, URL filtering, and cross-page content deduplication.',
-                inputSchema: CRAWL_TOOL_SCHEMA,
-            },
-            {
-                name: 'cache',
-                description: 'Query the local knowledge base of previously fetched content. ' +
-                    'Search cached pages by full-text query, URL pattern, or date. ' +
-                    'Can also return cache statistics or clear entries.',
-                inputSchema: CACHE_TOOL_SCHEMA,
-            },
-            {
-                name: 'extract',
-                description: 'Extract structured data from a web page. Supports CSS selector extraction, ' +
-                    'table-to-JSON conversion, metadata extraction (title, author, date, JSON-LD), ' +
-                    'and schema-based extraction (provide a JSON Schema to heuristically extract matching fields).',
-                inputSchema: EXTRACT_TOOL_SCHEMA,
-            },
-        ],
-    }));
-    server.setRequestHandler(CallToolRequestSchema, async (request) => {
-        const { name, arguments: args } = request.params;
-        if (name === 'fetch') {
-            const input = (args ?? {});
-            const result = await handleFetch(input, router);
-            return {
-                content: [{ type: 'text', text: JSON.stringify(result, null, 2) }],
-                isError: !!result.error,
-            };
-        }
-        if (name === 'search') {
-            const input = (args ?? {});
-            const result = await handleSearch(input, searchEngines, router);
-            return {
-                content: [{ type: 'text', text: JSON.stringify(result, null, 2) }],
-                isError: !!result.error,
-            };
-        }
-        if (name === 'crawl') {
-            const input = (args ?? {});
-            const result = await handleCrawl(input, router);
-            return {
-                content: [{ type: 'text', text: JSON.stringify(result, null, 2) }],
-                isError: !!result.error,
-            };
-        }
-        if (name === 'cache') {
-            const input = (args ?? {});
-            const result = handleCache(input);
-            return {
-                content: [{ type: 'text', text: JSON.stringify(result, null, 2) }],
-                isError: !!result.error,
-            };
-        }
-        if (name === 'extract') {
-            const input = (args ?? {});
-            const result = await handleExtract(input, router);
-            return {
-                content: [{ type: 'text', text: JSON.stringify(result, null, 2) }],
-                isError: !!result.error,
-            };
+  }
+  async function shutdown() {
+    log.info("Shutting down");
+    if (searxngBootstrap) {
+      await Promise.race([
+        searxngBootstrap.catch(() => {
+        }),
+        new Promise((r) => setTimeout(r, 2e3))
+      ]);
+    }
+    if (searxngProcess) await searxngProcess.stop();
+    if (dockerSearxng) await dockerSearxng.stop();
+    await browserPool.shutdown();
+    await closeDaemonBrowser().catch((e) => log.debug("closeDaemonBrowser failed", { error: e instanceof Error ? e.message : String(e) }));
+    resetEmbeddingService();
+    closeDatabase();
+  }
+  return {
+    searchEngines,
+    browserPool,
+    router,
+    backendStatus,
+    pluginRegistry,
+    shutdown,
+    bootstrapSearxng: () => {
+      searxngBootstrap = bootstrapSearxng();
+      return searxngBootstrap;
+    }
+  };
+}
+function createMcpServer(subsystems) {
+  const { searchEngines, router, backendStatus } = subsystems;
+  const server = new Server(
+    { name: "wigolo", version: SERVER_VERSION },
+    {
+      capabilities: { tools: {}, resources: {} },
+      instructions: WIGOLO_INSTRUCTIONS
+    }
+  );
+  server.setRequestHandler(ListResourcesRequestSchema, async () => ({
+    resources: [
+      {
+        uri: WIGOLO_DOCS_URI,
+        name: "Wigolo usage guide",
+        description: "Routing tables, performance budgets, auth flows, and other detail trimmed from the per-session instructions.",
+        mimeType: "text/markdown"
+      }
+    ]
+  }));
+  server.setRequestHandler(ReadResourceRequestSchema, async (request) => {
+    if (request.params.uri !== WIGOLO_DOCS_URI) {
+      throw new Error(`Unknown resource: ${request.params.uri}`);
+    }
+    return {
+      contents: [
+        {
+          uri: WIGOLO_DOCS_URI,
+          mimeType: "text/markdown",
+          text: WIGOLO_INSTRUCTIONS_FULL
         }
+      ]
+    };
+  });
+  server.setRequestHandler(ListToolsRequestSchema, async () => ({
+    tools: [
+      {
+        name: "fetch",
+        description: TOOL_DESCRIPTIONS.fetch,
+        inputSchema: FETCH_TOOL_SCHEMA
+      },
+      {
+        name: "search",
+        description: TOOL_DESCRIPTIONS.search,
+        inputSchema: SEARCH_TOOL_SCHEMA
+      },
+      {
+        name: "crawl",
+        description: TOOL_DESCRIPTIONS.crawl,
+        inputSchema: CRAWL_TOOL_SCHEMA
+      },
+      {
+        name: "cache",
+        description: TOOL_DESCRIPTIONS.cache,
+        inputSchema: CACHE_TOOL_SCHEMA
+      },
+      {
+        name: "extract",
+        description: TOOL_DESCRIPTIONS.extract,
+        inputSchema: EXTRACT_TOOL_SCHEMA
+      },
+      {
+        name: "find_similar",
+        description: TOOL_DESCRIPTIONS.find_similar,
+        inputSchema: FIND_SIMILAR_TOOL_SCHEMA
+      },
+      {
+        name: "research",
+        description: TOOL_DESCRIPTIONS.research,
+        inputSchema: RESEARCH_TOOL_SCHEMA
+      },
+      {
+        name: "agent",
+        description: TOOL_DESCRIPTIONS.agent,
+        inputSchema: AGENT_TOOL_SCHEMA
+      }
+    ]
+  }));
+  server.setRequestHandler(CallToolRequestSchema, async (request, extra) => {
+    const { name, arguments: args } = request.params;
+    const meta = request.params._meta;
+    const progressToken = meta?.progressToken;
+    const onProgress = progressToken !== void 0 && extra && typeof extra.sendNotification === "function" ? async (update) => {
+      try {
+        await extra.sendNotification({
+          method: "notifications/progress",
+          params: {
+            progressToken,
+            progress: update.progress,
+            total: update.total,
+            message: update.message
+          }
+        });
+      } catch (err) {
+        log.debug("sendNotification failed", { error: String(err) });
+      }
+    } : void 0;
+    if (name === "fetch") {
+      const input = args ?? {};
+      const r = await handleFetch(input, router);
+      if (!r.ok) {
         return {
-            content: [{ type: 'text', text: `Unknown tool: ${name}` }],
-            isError: true,
+          content: [{ type: "text", text: JSON.stringify({ error: r.error, error_reason: r.error_reason, stage: r.stage, ...r.hint ? { hint: r.hint } : {} }, null, 2) }],
+          isError: true
         };
-    });
-    const transport = new StdioServerTransport();
-    await server.connect(transport);
-    log.info('MCP server started');
-    const shutdown = async () => {
-        log.info('Shutting down');
-        if (searxngProcess)
-            await searxngProcess.stop();
-        if (dockerSearxng)
-            await dockerSearxng.stop();
-        await browserPool.shutdown();
-        closeDatabase();
-        await server.close();
-        process.exit(0);
+      }
+      return {
+        content: [{ type: "text", text: JSON.stringify(r.data, null, 2) }],
+        isError: false
+      };
+    }
+    if (name === "search") {
+      const input = args ?? {};
+      const samplingServer = server;
+      const r = await handleSearch(input, searchEngines, router, backendStatus, samplingServer, onProgress);
+      if (!r.ok) {
+        return {
+          content: [{ type: "text", text: JSON.stringify({ error: r.error, error_reason: r.error_reason, stage: r.stage, ...r.hint ? { hint: r.hint } : {} }, null, 2) }],
+          isError: true
+        };
+      }
+      const blocks = [];
+      if (r.data.warning) {
+        blocks.push({ type: "text", text: `[wigolo notice] ${r.data.warning}` });
+      }
+      blocks.push({ type: "text", text: JSON.stringify(r.data, null, 2) });
+      return {
+        content: blocks,
+        isError: !!r.data.error
+      };
+    }
+    if (name === "crawl") {
+      const input = args ?? {};
+      const result = await handleCrawl(input, router);
+      return {
+        content: [{ type: "text", text: JSON.stringify(result, null, 2) }],
+        isError: !!result.error
+      };
+    }
+    if (name === "cache") {
+      const input = args ?? {};
+      const result = await handleCache(input, router);
+      return {
+        content: [{ type: "text", text: JSON.stringify(result, null, 2) }],
+        isError: !!result.error
+      };
+    }
+    if (name === "extract") {
+      const input = args ?? {};
+      const r = await handleExtract(input, router);
+      if (!r.ok) {
+        return {
+          content: [{ type: "text", text: JSON.stringify({ error: r.error, error_reason: r.error_reason, stage: r.stage, ...r.hint ? { hint: r.hint } : {} }, null, 2) }],
+          isError: true
+        };
+      }
+      return {
+        content: [{ type: "text", text: JSON.stringify(r.data, null, 2) }],
+        isError: false
+      };
+    }
+    if (name === "find_similar") {
+      const input = args ?? {};
+      const r = await handleFindSimilar(input, searchEngines, router, backendStatus);
+      if (!r.ok) {
+        return {
+          content: [{ type: "text", text: JSON.stringify({ error: r.error, error_reason: r.error_reason, stage: r.stage, ...r.hint ? { hint: r.hint } : {} }, null, 2) }],
+          isError: true
+        };
+      }
+      return {
+        content: [{ type: "text", text: JSON.stringify(r.data, null, 2) }],
+        isError: false
+      };
+    }
+    if (name === "research") {
+      const input = args ?? {};
+      const samplingServer = server;
+      const r = await handleResearch(input, searchEngines, router, backendStatus, samplingServer);
+      if (!r.ok) {
+        return {
+          content: [{ type: "text", text: JSON.stringify({ error: r.error, error_reason: r.error_reason, stage: r.stage, ...r.hint ? { hint: r.hint } : {} }, null, 2) }],
+          isError: true
+        };
+      }
+      return {
+        content: [{ type: "text", text: JSON.stringify(r.data, null, 2) }],
+        isError: false
+      };
+    }
+    if (name === "agent") {
+      const input = args ?? {};
+      const samplingServer = server;
+      const r = await handleAgent(input, searchEngines, router, backendStatus, samplingServer);
+      if (!r.ok) {
+        return {
+          content: [{ type: "text", text: JSON.stringify({ error: r.error, error_reason: r.error_reason, stage: r.stage, ...r.hint ? { hint: r.hint } : {} }, null, 2) }],
+          isError: true
+        };
+      }
+      return {
+        content: [{ type: "text", text: JSON.stringify(r.data, null, 2) }],
+        isError: false
+      };
+    }
+    return {
+      content: [{ type: "text", text: `Unknown tool: ${name}` }],
+      isError: true
     };
-    process.on('SIGINT', shutdown);
-    process.on('SIGTERM', shutdown);
+  });
+  return server;
+}
+async function startServer() {
+  const subs = await initSubsystems();
+  const server = createMcpServer(subs);
+  const transport = new StdioServerTransport();
+  await server.connect(transport);
+  log.info("MCP server started");
+  maybeEagerWarmup();
+  subs.bootstrapSearxng().catch((err) => {
+    log.warn("search engine bootstrap failed", { error: String(err) });
+  });
+  const shutdown = async () => {
+    await subs.shutdown();
+    await server.close();
+    process.exit(0);
+  };
+  process.on("SIGINT", shutdown);
+  process.on("SIGTERM", shutdown);
 }
+export {
+  createMcpServer,
+  initSubsystems,
+  startServer
+};
 //# sourceMappingURL=server.js.map