@crawlbrulee/sdk 0.6.0 โ†’ 0.7.1

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.
package/README.md CHANGED
@@ -1,17 +1,28 @@
1
- # @crawlbrulee/sdk
1
+ # ๐Ÿฎ crawlbrulee js/ts sdk
2
2
 
3
- The official TypeScript / JavaScript SDK for the [crawlbrulee](https://crawlbrulee.com) web-scraping API.
3
+ [![npm](https://img.shields.io/npm/v/@crawlbrulee/sdk?style=flat-square&label=npm)](https://www.npmjs.com/package/@crawlbrulee/sdk)
4
+ [![types](https://img.shields.io/npm/types/@crawlbrulee/sdk?style=flat-square&label=types)](https://www.npmjs.com/package/@crawlbrulee/sdk)
5
+ [![license](https://img.shields.io/npm/l/@crawlbrulee/sdk?style=flat-square&label=license)](./LICENSE)
4
6
 
5
- - Fully typed.
7
+ the official js/ts sdk for the [crawlbrulee](https://crawlbrulee.com) web-scraping api โ€” published to npm as [`@crawlbrulee/sdk`](https://www.npmjs.com/package/@crawlbrulee/sdk). you
8
+ send a url, you get back markdown, cleaned html, links, images, metadata, or a screenshot.
9
+
10
+ - fully typed.
6
11
  - ESM + CommonJS, ships its own `.d.ts`.
7
- - Zero runtime dependencies โ€” just `fetch`.
8
- - Works on Node.js 22+, modern Deno, Bun, and runtimes where `fetch` is available.
12
+ - zero runtime dependencies โ€” just `fetch`.
13
+ - works on Node.js 22+, modern Deno, Bun, and runtimes where `fetch` is available.
14
+
15
+ this readme covers the sdk itself โ€” the client, the types, and the js-side ergonomics. for how
16
+ the api behaves โ€” endpoints, parameters, and error semantics โ€” please see our
17
+ [api docs](https://crawlbrulee.com/docs).
9
18
 
10
- > **Status:** v0.6.0 (beta). API surface is stabilizing โ€” expect minor breaking changes between 0.x releases.
19
+ > **status:** v0.7.0 (beta). the api surface is stabilizing โ€” expect minor breaking changes between 0.x releases.
20
+
21
+ **get a free api key** โ†’ [dashboard.crawlbrulee.com](https://dashboard.crawlbrulee.com)
11
22
 
12
23
  ---
13
24
 
14
- ## Install
25
+ ## install
15
26
 
16
27
  ```bash
17
28
  pnpm add @crawlbrulee/sdk
@@ -21,7 +32,7 @@ npm install @crawlbrulee/sdk
21
32
  yarn add @crawlbrulee/sdk
22
33
  ```
23
34
 
24
- ## Quickstart
35
+ ## quickstart
25
36
 
26
37
  ```ts
27
38
  import { Crawlbrulee } from '@crawlbrulee/sdk'
@@ -41,22 +52,38 @@ console.log(page.metadata?.title) // structured <head> metadata
41
52
  console.log(page.response_meta.usage.credits, 'credits charged') // usage accounting
42
53
  ```
43
54
 
44
- ### Configuration
55
+ ### authentication
56
+
57
+ every request carries your api key as `Authorization: Bearer <key>`. give it to the sdk one of two ways:
45
58
 
46
- | Option | Default | Description |
47
- | ----------- | ---------------- | ------------------------------------------------------------------------------------------ |
48
- | `apiKey` | โ€” | API key, sent as `Authorization: Bearer โ€ฆ`. **Required** โ€” or use `Crawlbrulee.fromEnv()`. |
49
- | `timeoutMs` | `0` (no timeout) | Per-request timeout (covers headers + body). A per-call `timeoutMs` overrides this. |
59
+ ```ts
60
+ // explicit โ€” pass the key directly
61
+ const crawlbrulee = new Crawlbrulee({ apiKey: 'cwbl_โ€ฆ' })
50
62
 
51
- `Crawlbrulee.fromEnv(overrides?)` reads the API key from `CRAWLBRULEE_API_KEY` and forwards any other option through `overrides`.
63
+ // from the environment โ€” reads CRAWLBRULEE_API_KEY
64
+ const crawlbrulee = Crawlbrulee.fromEnv()
65
+ ```
66
+
67
+ `Crawlbrulee.fromEnv(overrides?)` reads the key from `CRAWLBRULEE_API_KEY` and forwards any other option through
68
+ `overrides` (e.g. `Crawlbrulee.fromEnv({ timeoutMs: 30_000 })`). it throws if the variable is unset or empty. keys are
69
+ minted in the dashboard; see [authentication](https://crawlbrulee.com/docs/authentication) for how the api consumes them.
70
+
71
+ ### configuration
72
+
73
+ | option | default | description |
74
+ | ----------- | ----------------------------- | --------------------------------------------------------------------------------------------------- |
75
+ | `apiKey` | โ€” | api key, sent as `Authorization: Bearer โ€ฆ`. **required** โ€” or use `Crawlbrulee.fromEnv()`. |
76
+ | `baseUrl` | `https://api.crawlbrulee.com` | override the target host (local dev / staging). trailing slashes are stripped. |
77
+ | `timeoutMs` | `0` (no timeout) | per-request timeout in milliseconds (covers headers + body). a per-call `timeoutMs` overrides this. |
52
78
 
53
79
  ---
54
80
 
55
- ## API reference
81
+ ## api reference
56
82
 
57
- All methods return a `Promise` that resolves to the parsed JSON response, or rejects with a [`CrawlbruleeError`](#errors) subclass.
83
+ all methods return a `Promise` that resolves to the parsed json response, or rejects with a [`CrawlbruleeError`](#errors)
84
+ subclass.
58
85
 
59
- Every method accepts an optional second argument with per-call overrides:
86
+ every method accepts an optional second argument with per-call overrides:
60
87
 
61
88
  ```ts
62
89
  crawlbrulee.scrape(request, {
@@ -65,11 +92,11 @@ crawlbrulee.scrape(request, {
65
92
  })
66
93
  ```
67
94
 
68
- ### Scraping
95
+ ### scraping
69
96
 
70
97
  #### `crawlbrulee.scrape(request, options?)`
71
98
 
72
- Synchronously scrape a URL. The request blocks until the server is done.
99
+ scrape a url synchronously. the request blocks until the server is done.
73
100
 
74
101
  ```ts
75
102
  const page = await crawlbrulee.scrape({
@@ -93,43 +120,62 @@ const page = await crawlbrulee.scrape({
93
120
  })
94
121
  ```
95
122
 
96
- The response carries the extracted content alongside structured `metadata` (the parsed `<head>` tags โ€” `title`, `description`, OG/Twitter fields, โ€ฆ) and a `response_meta` envelope:
123
+ the response carries the extracted content alongside structured `metadata` (the parsed `<head>` tags โ€” `title`,
124
+ `description`, OG/Twitter fields, โ€ฆ) and a `response_meta` envelope:
97
125
 
98
126
  ```ts
99
127
  page.metadata?.title // structured <head> metadata (when extract.metadata, on by default)
100
128
 
101
129
  page.response_meta.usage.credits // credits charged โ€” 0 on a cache hit
102
- page.response_meta.usage.proxy // the resolved proxy tier actually used: 'none' | 'basic' | 'advanced' (never 'auto')
130
+ page.response_meta.usage.proxy // the resolved proxy tier actually used: 'basic' | 'advanced' (never 'auto')
103
131
  page.response_meta.usage.cache_hit // true when the result was served from cache
104
132
  ```
105
133
 
106
- Notes:
107
-
108
- - **`proxy` defaults to `auto`** when omitted โ€” it starts at the basic tier and escalates to advanced on failure, billed at the delivered tier. Pass `'basic'`, `'advanced'`, or `'none'` to pin a tier.
109
- - **Screenshots.** In rare cases a screenshot can't be captured; when that happens the rest of your requested outputs are still returned and the `screenshot` field is simply left out, so guard for it (`page.screenshot?.url`). Custom `viewport.width`/`height` are integers in `[16, 10000]` and `device_scale_factor` is in `[1, 4]`; out-of-range values are rejected with a `400`.
110
- - **`extract.images`** URLs preserve their query string and resolve document-relative `src`s against the full page URL (browser parity) โ€” the same rules as `links`.
111
-
112
- See [`ScrapeRequest`](src/types/scrape.ts) and [`ScrapeResponse`](src/types/scrape.ts) for every field, with inline documentation.
134
+ notes:
135
+
136
+ - **`proxy`**: defaults to `auto` when omitted โ€” it starts at the basic tier and escalates to advanced on failure,
137
+ billed at the delivered tier. pass `'basic'` or `'advanced'` to pin a tier. on the response,
138
+ `response_meta.usage.proxy` reports the tier we resolved and used โ€” never `'auto'`. see
139
+ [proxies & location](https://crawlbrulee.com/docs/proxies) for what each tier does.
140
+ - **`screenshot`**: custom `viewport.width`/`height` are integers in `[16, 10000]` and `device_scale_factor` is in
141
+ `[1, 4]`; out-of-range values are rejected with a `400`. full capture options:
142
+ [screenshots](https://crawlbrulee.com/docs/scrape/screenshots).
143
+ - **`extract.images`**: urls preserve their query string and resolve document-relative `src`s against the full page url
144
+ (browser parity) โ€” the same rules as `links`. every extract field is documented under
145
+ [extraction](https://crawlbrulee.com/docs/scrape/extraction).
146
+ - **`warnings`**: when we complete a scrape but something is worth flagging โ€” e.g. `screenshot_truncated` when a long
147
+ page exceeded the scrolling-screenshot height cap โ€” the codes land on `page.warnings`. they're stable, so you can switch
148
+ on them. fresh scrapes only; cache hits omit warnings.
149
+ - **`unsupported_fields`**: if you request an extract that doesn't apply to the content type (e.g. `markdown` of a pdf),
150
+ that field name comes back on `page.unsupported_fields` and the rest of your payload is still returned.
151
+
152
+ see [`ScrapeRequest`](src/types/scrape.ts) and [`ScrapeResponse`](src/types/scrape.ts) for every field, with inline
153
+ documentation โ€” and the [scrape endpoint](https://crawlbrulee.com/docs/scrape) reference for the api-side contract those
154
+ types mirror.
113
155
 
114
156
  #### `crawlbrulee.scrapeAsync(request, options?)`
115
157
 
116
- Submit a scrape job in the background. Returns immediately with a `job_id`.
158
+ submit a scrape job in the background. returns immediately with a `job_id`.
117
159
 
118
160
  ```ts
119
161
  const { job_id } = await crawlbrulee.scrapeAsync({ url: 'https://example.com' })
120
162
  ```
121
163
 
164
+ pass a `webhook` to be notified on completion instead of polling โ€” see [webhooks](#webhooks).
165
+
122
166
  #### `crawlbrulee.getScrapeStatus(jobId, options?)`
123
167
 
124
- Look up the current status of an async job โ€” `pending`, `running`, `done`, or `failed`.
168
+ look up the current state of an async job โ€” `pending`, `running`, `done`, or `failed`. the response carries `job_id` and
169
+ `created_at` (snake_case, straight off the wire). once the job is `done`, it also carries usage accounting on
170
+ `response_meta.usage` (`credits`, `proxy`, `cache_hit`).
125
171
 
126
172
  #### `crawlbrulee.getScrapeResult(jobId, options?)`
127
173
 
128
- Fetch the result of a completed async job. Throws if the job hasn't finished yet.
174
+ fetch the result of a completed async job. throws if the job hasn't finished yet.
129
175
 
130
176
  #### `crawlbrulee.waitForScrape(jobId, options?)`
131
177
 
132
- Poll an async job until it reaches a terminal state, then return the scrape result.
178
+ poll an async job until it reaches a terminal state, then return the scrape result.
133
179
 
134
180
  ```ts
135
181
  const { job_id } = await crawlbrulee.scrapeAsync({ url: 'https://example.com' })
@@ -140,13 +186,16 @@ const page = await crawlbrulee.waitForScrape(job_id, {
140
186
  })
141
187
  ```
142
188
 
143
- Throws a `CrawlbruleeError` with `errorName: 'job_failed'` if the job ends in `failed`, or `errorName: 'request_timeout'` if the wait expires.
189
+ throws a `CrawlbruleeError` with `errorName: 'job_failed'` if the job ends in `failed`, or `errorName: 'request_timeout'`
190
+ if the wait expires. the job lifecycle itself โ€” states, retention, and when to prefer async over sync โ€” is documented
191
+ under [async scrape](https://crawlbrulee.com/docs/scrape/async).
144
192
 
145
- ### Mapping
193
+ ### mapping
146
194
 
147
195
  #### `crawlbrulee.map(request, options?)`
148
196
 
149
- Build (or return a cached) link map for a website. Combines sitemap discovery with the freshest cached homepage scrape when available.
197
+ build (or return a cached) link map for a website. combines sitemap discovery with the freshest cached homepage scrape
198
+ when available.
150
199
 
151
200
  ```ts
152
201
  const result = await crawlbrulee.map({
@@ -162,25 +211,41 @@ console.log(result.links.length, 'urls on page 1 of', result.response_meta.pagin
162
211
  console.log(result.response_meta.usage.credits, 'credits charged') // usage accounting, alongside pagination + truncation
163
212
  ```
164
213
 
165
- ### Account
214
+ `result.response_meta` carries `usage` (credits / resolved `proxy` / `cache_hit`) alongside the map-specific `pagination`
215
+ and `truncation` blocks. see the [map endpoint](https://crawlbrulee.com/docs/map) for discovery rules and pagination
216
+ semantics.
217
+
218
+ ### account
166
219
 
167
220
  #### `crawlbrulee.usage(options?)`
168
221
 
169
- Return the current billing-cycle snapshot โ€” total/used/available credits, used quota percentage, max concurrency, and the cycle reset timestamp.
222
+ return the current billing-cycle snapshot โ€” `total_credits`, `used_credits`, `available_credits`, `used_quota_percent`,
223
+ `max_concurrency`, and the `usage_reset` timestamp.
170
224
 
171
225
  #### `crawlbrulee.whoami(options?)`
172
226
 
173
- Return the organization name and identifying details of the API token used to authenticate the request.
227
+ return the organization name and token identity behind the api key (`organization_name`, `token_name`, and a
228
+ safe-to-display `token_preview`). use it to confirm which key is in play before a destructive operation.
229
+
230
+ what a call costs, and how credits are counted, is documented under
231
+ [credits & pricing](https://crawlbrulee.com/docs/credits-and-pricing).
174
232
 
175
233
  ---
176
234
 
177
- ## Webhooks
235
+ ## webhooks
236
+
237
+ when an async scrape job finishes, crawlbrulee can `POST` a `scrape.complete` webhook to your configured endpoint. the
238
+ sdk ships two helpers for it.
178
239
 
179
- When an async scrape job finishes, crawlbrulee can `POST` a `scrape.complete` webhook to your configured endpoint. The SDK ships two helpers for it.
240
+ the delivery contract and payload shape live under [webhooks](https://crawlbrulee.com/docs/scrape/webhooks); the
241
+ signature scheme is specified in [webhook verification](https://crawlbrulee.com/docs/webhook-verification). what follows
242
+ is how this sdk helps you consume them.
180
243
 
181
- ### Triggering a webhook (`scrapeAsync`)
244
+ ### triggering a webhook (`scrapeAsync`)
182
245
 
183
- Pass a `webhook` to `scrapeAsync` to have the API deliver a single signed `scrape.complete` `POST` when the job reaches a terminal state. This is **async-only** โ€” the synchronous `scrape()` response _is_ the notification, so it does not accept a `webhook`.
246
+ pass a `webhook` to `scrapeAsync` to have us deliver a single signed `scrape.complete` `POST` when the job reaches a
247
+ terminal state. this is **async-only** โ€” the synchronous `scrape()` response _is_ the notification, so it does not accept
248
+ a `webhook`.
184
249
 
185
250
  ```ts
186
251
  const { job_id } = await crawlbrulee.scrapeAsync({
@@ -195,11 +260,16 @@ const { job_id } = await crawlbrulee.scrapeAsync({
195
260
  })
196
261
  ```
197
262
 
198
- Configure the signing secret used for these deliveries in the dashboard (**Account โ†’ Webhooks**); there is no per-request secret. When the delivery arrives, verify it with [`verifyWebhookSignature`](#verifywebhooksignatureoptions) and read your `metadata` back from `webhook.data.metadata`. The delivery also carries usage accounting on `webhook.data.response_meta.usage` (`credits`, `proxy`, `cache_hit`). See [`AsyncScrapeWebhook`](src/types/scrape.ts) for the full field documentation.
263
+ configure the signing secret used for these deliveries in the dashboard (**account โ†’ webhooks**). there is no per-request
264
+ secret - when the delivery arrives, verify it with [`verifyWebhookSignature`](#verifywebhooksignatureoptions) and read your `metadata` back from
265
+ `webhook.data.metadata`. the delivery also carries usage accounting on `webhook.data.response_meta.usage` (`credits`,
266
+ `proxy`, `cache_hit`). see [`AsyncScrapeWebhook`](src/types/scrape.ts) for the full field documentation.
199
267
 
200
268
  ### `verifyWebhookSignature(options)`
201
269
 
202
- A standalone, network-free helper (built on Web Crypto, so it runs on Node.js 22+, browsers, Bun, Deno, and edge) that verifies the `X-Cwbl-Signature` header. **It returns a result object rather than throwing** โ€” a failed verification is normal control flow.
270
+ a standalone, network-free helper (built on Web Crypto, so it runs on Node.js 22+, browsers, Bun, Deno, and edge) that
271
+ verifies the `X-Cwbl-Signature` header. **it returns a result object rather than throwing** โ€” a failed verification is
272
+ normal control flow.
203
273
 
204
274
  ```ts
205
275
  import { verifyWebhookSignature } from '@crawlbrulee/sdk'
@@ -218,11 +288,15 @@ if (result.verified) {
218
288
  }
219
289
  ```
220
290
 
221
- During a **signing-secret rotation grace window** the API sends a second `X-Cwbl-Signature-Rotated` header signed with the previous secret. `verifyWebhookSignature` tries your `secret` against the primary header first, then the rotated one, and reports which matched via `signedWith` โ€” so verification keeps working whether you still hold the old secret or have already rotated to the new one.
291
+ during a **signing-secret rotation grace window** we send a second `X-Cwbl-Signature-Rotated` header signed with the
292
+ previous secret. `verifyWebhookSignature` tries your `secret` against the primary header first, then the rotated one,
293
+ and reports which matched via `signedWith` โ€” so verification keeps working whether you still hold the old secret or have
294
+ already rotated to the new one.
222
295
 
223
296
  ### `crawlbrulee.fetchScrapeResultFromWebhook(webhook, options?)`
224
297
 
225
- Given a verified `scrape.complete` body, fetch the scrape result. Returns `getScrapeResult(job_id)` for a `success` job; throws a `CrawlbruleeError` for `failed` (carrying the failure message) or `cancelled` jobs.
298
+ given a verified `scrape.complete` body, fetch the scrape result. returns `getScrapeResult(job_id)` for a `success` job;
299
+ throws a `CrawlbruleeError` for `failed` (carrying the failure message) or `cancelled` jobs.
226
300
 
227
301
  ```ts
228
302
  import { Crawlbrulee, verifyWebhookSignature, type ScrapeCompleteWebhook } from '@crawlbrulee/sdk'
@@ -246,23 +320,25 @@ app.post('/webhooks/crawlbrulee', async (req, res) => {
246
320
  })
247
321
  ```
248
322
 
249
- Always verify the signature **before** parsing or trusting the body. The `X-Cwbl-Event-Id` header (also `webhook.event_id`) is a stable id you can use to de-duplicate deliveries.
323
+ always verify the signature **before** parsing or trusting the body. the `X-Cwbl-Event-Id` header (also
324
+ `webhook.event_id`) is a stable id you can use to de-duplicate deliveries.
250
325
 
251
326
  ---
252
327
 
253
- ## Errors
328
+ ## errors
254
329
 
255
- Every failure raised by the SDK extends [`CrawlbruleeError`](src/errors.ts). Typed subclasses are exported for the most actionable cases:
330
+ every failure raised by the sdk extends [`CrawlbruleeError`](src/errors.ts). typed subclasses are exported for the most actionable
331
+ cases:
256
332
 
257
- | Class | When it's raised |
333
+ | class | when it's raised |
258
334
  | ---------------------- | ---------------------------------------------------------------------------------------------------- |
259
- | `AuthenticationError` | 401 / 403 responses (missing, invalid, or unauthorized API key). |
260
- | `RateLimitError` | 429 responses. Exposes `retryAfterMs` and `limitedBy` when the server provided them. |
261
- | `UsageAllocationError` | The org's plan limit was hit. Exposes `reason` (`credit_limit`, `concurrency_limit`, โ€ฆ) and `usage`. |
335
+ | `AuthenticationError` | 401 / 403 responses (missing, invalid, or unauthorized api key). |
336
+ | `RateLimitError` | 429 responses. exposes `retryAfterMs` and `limitedBy` when the server provided them. |
337
+ | `UsageAllocationError` | the org's plan limit was hit. exposes `reason` (`credit_limit`, `concurrency_limit`, โ€ฆ) and `usage`. |
262
338
  | `ValidationError` | 4xx caused by a bad request (`invalid_url`, `url_too_long`, `blocked_url`, โ€ฆ). |
263
339
  | `NotFoundError` | 404 responses (e.g. unknown async `jobId`). |
264
- | `TransportError` | Network failures, aborts, non-JSON responses, request body read failures. |
265
- | `CrawlbruleeError` | Base class โ€” used for any other API error. Always has `status`, `errorName`, `message`. |
340
+ | `TransportError` | network failures, aborts, non-json responses, request body read failures. |
341
+ | `CrawlbruleeError` | base class โ€” used for any other api error. always has `status`, `errorName`, `message`. |
266
342
 
267
343
  ```ts
268
344
  import { Crawlbrulee, RateLimitError, UsageAllocationError } from '@crawlbrulee/sdk'
@@ -282,26 +358,13 @@ try {
282
358
  }
283
359
  ```
284
360
 
285
- For exhaustive branching, switch on `err.errorName` โ€” the literal-typed union is exported as `ApiErrorName`.
286
-
287
- ### Rate limits
288
-
289
- Limits are per-plan, with **separate buckets for sync and async** (requests per minute). Synchronous `scrape()` **and** `map()` count against the sync bucket; `scrapeAsync()` submissions have their own bucket.
290
-
291
- | Plan | Sync (rpm) | Async (rpm) |
292
- | -------- | ---------- | ----------- |
293
- | Free | 50 | 100 |
294
- | Starter | 100 | 300 |
295
- | Pro | 350 | 1000 |
296
- | Advanced | 1000 | 3000 |
297
-
298
- When you exceed a bucket the API returns `429` and the SDK throws a `RateLimitError` โ€” read `retryAfterMs` to back off.
299
-
300
- ---
361
+ for exhaustive branching, switch on `err.errorName` โ€” the literal-typed union is exported as `ApiErrorName`. the
362
+ `isCrawlbruleeError(err)` type guard narrows an `unknown` to the base error. the api docs carry the canonical
363
+ [error reference](https://crawlbrulee.com/docs/errors) โ€” every `errorName`, what causes it, and how to recover.
301
364
 
302
- ## Cancellation and timeouts
365
+ ## cancellation and timeouts
303
366
 
304
- Every method accepts an `AbortSignal`:
367
+ every method accepts an `AbortSignal`:
305
368
 
306
369
  ```ts
307
370
  const controller = new AbortController()
@@ -310,11 +373,12 @@ const page = crawlbrulee.scrape({ url: 'https://slow.example.com' }, { signal: c
310
373
  setTimeout(() => controller.abort(), 5_000)
311
374
  ```
312
375
 
313
- The per-call `timeoutMs` and the caller's signal are composed โ€” whichever fires first wins.
376
+ the per-call `timeoutMs` and the caller's signal are composed โ€” whichever fires first wins. a fired timeout surfaces as
377
+ a `TransportError` with `errorName: 'request_timeout'`; an aborted signal as `errorName: 'client_closed_request'`.
314
378
 
315
379
  ---
316
380
 
317
- ## Development
381
+ ## development
318
382
 
319
383
  ```bash
320
384
  pnpm install
@@ -324,4 +388,20 @@ pnpm lint # eslint
324
388
  pnpm build # tsup โ†’ dist/
325
389
  ```
326
390
 
327
- The SDK has zero runtime dependencies on purpose. Please keep it that way when contributing.
391
+ the sdk has zero runtime dependencies on purpose. please keep it that way when contributing.
392
+
393
+ ## part of the crawlbrulee toolkit
394
+
395
+ one api, many ways to call it:
396
+
397
+ - **[js/ts sdk](https://github.com/crawlbrulee/crawlbrulee-js)** โ€” `@crawlbrulee/sdk` (this one)
398
+ - **[python sdk](https://github.com/crawlbrulee/crawlbrulee-py)** โ€” `crawlbrulee` on pypi
399
+ - **[cli](https://github.com/crawlbrulee/crawlbrulee-cli)** โ€” `npx crawlbrulee`
400
+ - **[mcp server](https://github.com/crawlbrulee/crawlbrulee-mcp)** โ€” `@crawlbrulee/mcp`, for ai agents
401
+ - **[agent skills](https://github.com/crawlbrulee/crawlbrulee-skills)** โ€” for skills-aware coding agents
402
+
403
+ docs: [crawlbrulee.com/docs](https://crawlbrulee.com/docs) ยท dashboard: [dashboard.crawlbrulee.com](https://dashboard.crawlbrulee.com)
404
+
405
+ ## license
406
+
407
+ [AGPL-3.0-only](./LICENSE)