Merge branch 'main' of https://github.com/mendableai/firecrawl

apps/api/src/controllers/v1/types.ts

@@ -30,7 +30,14 @@ export const url = z.preprocess(
       "URL must have a valid top-level domain or be a valid path"
     )
     .refine(
-      (x) => checkUrl(x as string),
+      (x) => {
+        try {
+          checkUrl(x as string)
+          return true;
+        } catch (_) {
+          return false;
+        }
+      },
       "Invalid URL"
     )
     .refine(

apps/api/src/index.ts

@@ -201,16 +201,20 @@ if (cluster.isMaster) {
   Sentry.setupExpressErrorHandler(app);
 
   app.use((err: unknown, req: Request<{}, ErrorResponse, undefined>, res: ResponseWithSentry<ErrorResponse>, next: NextFunction) => {
+    if (err instanceof SyntaxError && 'status' in err && err.status === 400 && 'body' in err) {
+      return res.status(400).json({ success: false, error: 'Bad request, malformed JSON' });
+    }
+
     const id = res.sentry ?? uuidv4();
     let verbose = JSON.stringify(err);
     if (verbose === "{}") {
-        if (err instanceof Error) {
-            verbose = JSON.stringify({
-                message: err.message,
-                name: err.name,
-                stack: err.stack,
-            });
-        }
+      if (err instanceof Error) {
+        verbose = JSON.stringify({
+          message: err.message,
+          name: err.name,
+          stack: err.stack,
+        });
+      }
     }
 
     Logger.error("Error occurred in request! (" + req.path + ") -- ID " + id  + " -- " + verbose);

apps/api/src/routes/v1.ts

@@ -83,7 +83,7 @@ function idempotencyMiddleware(req: Request, res: Response, next: NextFunction)
 }
 
 function blocklistMiddleware(req: Request, res: Response, next: NextFunction) {
-    if (req.body.url && isUrlBlocked(req.body.url)) {
+    if (typeof req.body.url === "string" && isUrlBlocked(req.body.url)) {
         if (!res.headersSent) {
             return res.status(403).json({ success: false, error: "URL is blocked. Firecrawl currently does not support social media scraping due to policy restrictions." });
         }

apps/api/src/scraper/WebScraper/scrapers/fireEngine.ts

@@ -55,7 +55,7 @@ export async function scrapWithFireEngine({
   try {
     const reqParams = await generateRequestParams(url);
     let waitParam = reqParams["params"]?.wait ?? waitFor;
-    let engineParam = reqParams["params"]?.engine ?? reqParams["params"]?.fireEngineOptions?.engine ?? fireEngineOptions?.engine  ?? "playwright";
+    let engineParam = reqParams["params"]?.engine ?? reqParams["params"]?.fireEngineOptions?.engine ?? fireEngineOptions?.engine  ?? "chrome-cdp";
     let screenshotParam = reqParams["params"]?.screenshot ?? screenshot;
     let fullPageScreenshotParam = reqParams["params"]?.fullPageScreenshot ?? fullPageScreenshot;
     let fireEngineOptionsParam : FireEngineOptions = reqParams["params"]?.fireEngineOptions ?? fireEngineOptions;

apps/api/src/scraper/WebScraper/single_url.ts

@@ -96,15 +96,15 @@ function getScrapingFallbackOrder(
     "fetch",
   ].filter(Boolean);
 
-  if (isWaitPresent || isScreenshotPresent || isHeadersPresent) {
-    defaultOrder = [
-      "fire-engine",
-      useFireEngine ? undefined : "playwright",
-      ...defaultOrder.filter(
-        (scraper) => scraper !== "fire-engine" && scraper !== "playwright"
-      ),
-    ].filter(Boolean);
-  }
+  // if (isWaitPresent || isScreenshotPresent || isHeadersPresent) {
+  //   defaultOrder = [
+  //     "fire-engine",
+  //     useFireEngine ? undefined : "playwright",
+  //     ...defaultOrder.filter(
+  //       (scraper) => scraper !== "fire-engine" && scraper !== "playwright"
+  //     ),
+  //   ].filter(Boolean);
+  // }
 
   const filteredDefaultOrder = defaultOrder.filter(
     (scraper: (typeof baseScrapers)[number]) =>

apps/api/src/services/rate-limiter.ts

@@ -6,7 +6,7 @@ const RATE_LIMITS = {
   crawl: {
     default: 3,
     free: 2,
-    starter: 3,
+    starter: 10,
     standard: 5,
     standardOld: 40,
     scale: 50,
@@ -19,9 +19,9 @@ const RATE_LIMITS = {
   scrape: {
     default: 20,
     free: 10,
-    starter: 20,
+    starter: 100,
     standard: 100,
-    standardOld: 40,
+    standardOld: 100,
     scale: 500,
     hobby: 20,
     standardNew: 100,
@@ -32,8 +32,8 @@ const RATE_LIMITS = {
   search: {
     default: 20,
     free: 5,
-    starter: 20,
-    standard: 40,
+    starter: 50,
+    standard: 50,
     standardOld: 40,
     scale: 500,
     hobby: 10,
@@ -45,9 +45,9 @@ const RATE_LIMITS = {
   map:{
     default: 20,
     free: 5,
-    starter: 20,
-    standard: 40,
-    standardOld: 40,
+    starter: 50,
+    standard: 50,
+    standardOld: 50,
     scale: 500,
     hobby: 10,
     standardNew: 50,