@mendable/firecrawl 4.13.1 → 4.14.0

This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
Files changed (9) hide show
  1. package/dist/{chunk-37UG5UXZ.js → chunk-BNFKCLH7.js} +1 -1
  2. package/dist/index.cjs +3 -1
  3. package/dist/index.d.cts +3 -0
  4. package/dist/index.d.ts +3 -0
  5. package/dist/index.js +4 -2
  6. package/dist/{package-JD6PB5ZT.js → package-WTVRF5QB.js} +1 -1
  7. package/package.json +1 -1
  8. package/src/v2/methods/crawl.ts +2 -0
  9. package/src/v2/types.ts +3 -1
package/dist/{chunk-37UG5UXZ.js → chunk-BNFKCLH7.js} RENAMED
@@ -8,7 +8,7 @@ var require_package = __commonJS({
8
8
  "package.json"(exports, module) {
9
9
  module.exports = {
10
10
  name: "@mendable/firecrawl-js",
11
- version: "4.13.1",
11
+ version: "4.14.0",
12
12
  description: "JavaScript SDK for Firecrawl API",
13
13
  main: "dist/index.js",
14
14
  types: "dist/index.d.ts",
package/dist/index.cjs CHANGED
@@ -35,7 +35,7 @@ var require_package = __commonJS({
35
35
  "package.json"(exports2, module2) {
36
36
  module2.exports = {
37
37
  name: "@mendable/firecrawl-js",
38
- version: "4.13.1",
38
+ version: "4.14.0",
39
39
  description: "JavaScript SDK for Firecrawl API",
40
40
  main: "dist/index.js",
41
41
  types: "dist/index.d.ts",
@@ -540,12 +540,14 @@ function prepareCrawlPayload(request) {
540
540
  if (request.maxDiscoveryDepth != null) data.maxDiscoveryDepth = request.maxDiscoveryDepth;
541
541
  if (request.sitemap != null) data.sitemap = request.sitemap;
542
542
  if (request.ignoreQueryParameters != null) data.ignoreQueryParameters = request.ignoreQueryParameters;
543
+ if (request.deduplicateSimilarURLs != null) data.deduplicateSimilarURLs = request.deduplicateSimilarURLs;
543
544
  if (request.limit != null) data.limit = request.limit;
544
545
  if (request.crawlEntireDomain != null) data.crawlEntireDomain = request.crawlEntireDomain;
545
546
  if (request.allowExternalLinks != null) data.allowExternalLinks = request.allowExternalLinks;
546
547
  if (request.allowSubdomains != null) data.allowSubdomains = request.allowSubdomains;
547
548
  if (request.delay != null) data.delay = request.delay;
548
549
  if (request.maxConcurrency != null) data.maxConcurrency = request.maxConcurrency;
550
+ if (request.regexOnFullURL != null) data.regexOnFullURL = request.regexOnFullURL;
549
551
  if (request.webhook != null) data.webhook = request.webhook;
550
552
  if (request.integration != null && request.integration.trim()) data.integration = request.integration.trim();
551
553
  if (request.scrapeOptions) {
package/dist/index.d.cts CHANGED
@@ -101,6 +101,7 @@ interface ScrapeOptions {
101
101
  mobile?: boolean;
102
102
  parsers?: Array<string | {
103
103
  type: 'pdf';
104
+ mode?: 'fast' | 'auto' | 'ocr';
104
105
  maxPages?: number;
105
106
  }>;
106
107
  actions?: ActionOption[];
@@ -382,6 +383,7 @@ interface CrawlOptions {
382
383
  maxDiscoveryDepth?: number | null;
383
384
  sitemap?: 'skip' | 'include' | 'only';
384
385
  ignoreQueryParameters?: boolean;
386
+ deduplicateSimilarURLs?: boolean;
385
387
  limit?: number | null;
386
388
  crawlEntireDomain?: boolean;
387
389
  allowExternalLinks?: boolean;
@@ -390,6 +392,7 @@ interface CrawlOptions {
390
392
  maxConcurrency?: number | null;
391
393
  webhook?: string | WebhookConfig | null;
392
394
  scrapeOptions?: ScrapeOptions | null;
395
+ regexOnFullURL?: boolean;
393
396
  zeroDataRetention?: boolean;
394
397
  integration?: string;
395
398
  }
package/dist/index.d.ts CHANGED
@@ -101,6 +101,7 @@ interface ScrapeOptions {
101
101
  mobile?: boolean;
102
102
  parsers?: Array<string | {
103
103
  type: 'pdf';
104
+ mode?: 'fast' | 'auto' | 'ocr';
104
105
  maxPages?: number;
105
106
  }>;
106
107
  actions?: ActionOption[];
@@ -382,6 +383,7 @@ interface CrawlOptions {
382
383
  maxDiscoveryDepth?: number | null;
383
384
  sitemap?: 'skip' | 'include' | 'only';
384
385
  ignoreQueryParameters?: boolean;
386
+ deduplicateSimilarURLs?: boolean;
385
387
  limit?: number | null;
386
388
  crawlEntireDomain?: boolean;
387
389
  allowExternalLinks?: boolean;
@@ -390,6 +392,7 @@ interface CrawlOptions {
390
392
  maxConcurrency?: number | null;
391
393
  webhook?: string | WebhookConfig | null;
392
394
  scrapeOptions?: ScrapeOptions | null;
395
+ regexOnFullURL?: boolean;
393
396
  zeroDataRetention?: boolean;
394
397
  integration?: string;
395
398
  }
package/dist/index.js CHANGED
@@ -1,6 +1,6 @@
1
1
  import {
2
2
  require_package
3
- } from "./chunk-37UG5UXZ.js";
3
+ } from "./chunk-BNFKCLH7.js";
4
4
 
5
5
  // src/v2/utils/httpClient.ts
6
6
  import axios from "axios";
@@ -423,12 +423,14 @@ function prepareCrawlPayload(request) {
423
423
  if (request.maxDiscoveryDepth != null) data.maxDiscoveryDepth = request.maxDiscoveryDepth;
424
424
  if (request.sitemap != null) data.sitemap = request.sitemap;
425
425
  if (request.ignoreQueryParameters != null) data.ignoreQueryParameters = request.ignoreQueryParameters;
426
+ if (request.deduplicateSimilarURLs != null) data.deduplicateSimilarURLs = request.deduplicateSimilarURLs;
426
427
  if (request.limit != null) data.limit = request.limit;
427
428
  if (request.crawlEntireDomain != null) data.crawlEntireDomain = request.crawlEntireDomain;
428
429
  if (request.allowExternalLinks != null) data.allowExternalLinks = request.allowExternalLinks;
429
430
  if (request.allowSubdomains != null) data.allowSubdomains = request.allowSubdomains;
430
431
  if (request.delay != null) data.delay = request.delay;
431
432
  if (request.maxConcurrency != null) data.maxConcurrency = request.maxConcurrency;
433
+ if (request.regexOnFullURL != null) data.regexOnFullURL = request.regexOnFullURL;
432
434
  if (request.webhook != null) data.webhook = request.webhook;
433
435
  if (request.integration != null && request.integration.trim()) data.integration = request.integration.trim();
434
436
  if (request.scrapeOptions) {
@@ -1471,7 +1473,7 @@ var FirecrawlApp = class {
1471
1473
  if (typeof process !== "undefined" && process.env && process.env.npm_package_version) {
1472
1474
  return process.env.npm_package_version;
1473
1475
  }
1474
- const packageJson = await import("./package-JD6PB5ZT.js");
1476
+ const packageJson = await import("./package-WTVRF5QB.js");
1475
1477
  return packageJson.default.version;
1476
1478
  } catch (error) {
1477
1479
  const isTest = typeof process !== "undefined" && (process.env.JEST_WORKER_ID != null || false);
package/dist/{package-JD6PB5ZT.js → package-WTVRF5QB.js} RENAMED
@@ -1,4 +1,4 @@
1
1
  import {
2
2
  require_package
3
- } from "./chunk-37UG5UXZ.js";
3
+ } from "./chunk-BNFKCLH7.js";
4
4
  export default require_package();
package/package.json CHANGED
@@ -1,6 +1,6 @@
1
1
  {
2
2
  "name": "@mendable/firecrawl",
3
- "version": "4.13.1",
3
+ "version": "4.14.0",
4
4
  "description": "JavaScript SDK for Firecrawl API",
5
5
  "main": "dist/index.js",
6
6
  "types": "dist/index.d.ts",
package/src/v2/methods/crawl.ts CHANGED
@@ -28,12 +28,14 @@ function prepareCrawlPayload(request: CrawlRequest): Record<string, unknown> {
28
28
  if (request.maxDiscoveryDepth != null) data.maxDiscoveryDepth = request.maxDiscoveryDepth;
29
29
  if (request.sitemap != null) data.sitemap = request.sitemap;
30
30
  if (request.ignoreQueryParameters != null) data.ignoreQueryParameters = request.ignoreQueryParameters;
31
+ if (request.deduplicateSimilarURLs != null) data.deduplicateSimilarURLs = request.deduplicateSimilarURLs;
31
32
  if (request.limit != null) data.limit = request.limit;
32
33
  if (request.crawlEntireDomain != null) data.crawlEntireDomain = request.crawlEntireDomain;
33
34
  if (request.allowExternalLinks != null) data.allowExternalLinks = request.allowExternalLinks;
34
35
  if (request.allowSubdomains != null) data.allowSubdomains = request.allowSubdomains;
35
36
  if (request.delay != null) data.delay = request.delay;
36
37
  if (request.maxConcurrency != null) data.maxConcurrency = request.maxConcurrency;
38
+ if (request.regexOnFullURL != null) data.regexOnFullURL = request.regexOnFullURL;
37
39
  if (request.webhook != null) data.webhook = request.webhook;
38
40
  if (request.integration != null && request.integration.trim()) data.integration = request.integration.trim();
39
41
  if (request.scrapeOptions) {
package/src/v2/types.ts CHANGED
@@ -145,7 +145,7 @@ export interface ScrapeOptions {
145
145
  timeout?: number;
146
146
  waitFor?: number;
147
147
  mobile?: boolean;
148
- parsers?: Array<string | { type: 'pdf'; maxPages?: number }>;
148
+ parsers?: Array<string | { type: 'pdf'; mode?: 'fast' | 'auto' | 'ocr'; maxPages?: number }>;
149
149
  actions?: ActionOption[];
150
150
  location?: LocationConfig;
151
151
  skipTlsVerification?: boolean;
@@ -462,6 +462,7 @@ export interface CrawlOptions {
462
462
  maxDiscoveryDepth?: number | null;
463
463
  sitemap?: 'skip' | 'include' | 'only';
464
464
  ignoreQueryParameters?: boolean;
465
+ deduplicateSimilarURLs?: boolean;
465
466
  limit?: number | null;
466
467
  crawlEntireDomain?: boolean;
467
468
  allowExternalLinks?: boolean;
@@ -470,6 +471,7 @@ export interface CrawlOptions {
470
471
  maxConcurrency?: number | null;
471
472
  webhook?: string | WebhookConfig | null;
472
473
  scrapeOptions?: ScrapeOptions | null;
474
+ regexOnFullURL?: boolean;
473
475
  zeroDataRetention?: boolean;
474
476
  integration?: string;
475
477
  }