Bright Data vs Crawl4AI

Bright Data 69

API keys carry one of five roles (Admin, Finance, Ops, Limit, User), can be given an expiry, and are revoked or refreshed in the control panel, and the hosted MCP also takes OAuth 2.1 with PKCE and one mcp scope (30). The README's default hosted setup still puts the key in the URL as ?token=, a documented option, so less 10 (20). User-role keys can't change configuration, tools load by group, and tools carry readOnlyHint and destructiveHint, but browser tools can fill forms and click with no confirmation step from the server (14 of 20). Since 3.1.0 results from search, scrape, scraper, dataset and browser tools are wrapped as untrusted web content with a per-call marker, and since 2.12.0 a content gate, documented in the README and on by default, asks the user through MCP elicitation before returning page text that looks like instructions to change the MCP configuration (12 of 15). The MCP session_stats tool reports calls in a session. We didn't find a per-request log for operators (5 of 15). ISO 27001, 27017 and 27018, SOC 2 Type II under NDA, a public SOC 3, a managed private bug bounty and security fixes in the MCP changelog per the 1 October check. No security.txt (404 on 4 November). OSV holds no advisory for @brightdata/mcp (18 of 20). 20+14+12+5+18 = 69.

Crawl4AI 46

Bearer keys, shown once and regenerable, with 3 to 30 keys per plan. claude.ai and ChatGPT sign in through a flow that advertises one scope, mcp, and each app shows as a revocable key. No per-key scopes, and the docs don't offer the key in a URL (22 of 30). A key can call billing routes, and the MCP topup and recharge tools charge the saved card, though llms.txt tells agents to ask first and accounts can set a spend cap. No read-only key (6 of 20). Scraped pages are untrusted text and no prompt-injection guidance was found in llms.txt or the API docs (0). Per-call cost, balance and fetch path in response headers and fields, no access log documented (6 of 15). The security page (4 October 2026) gives a disclosure process with fix targets and says SOC 2 is in progress, and the library publishes GitHub advisories with reporter credits. security.txt redirects to a 404, and no bounty found (12 of 20). 22 + 6 + 0 + 6 + 12 = 46.

Bright Data 74

No OpenAPI file found for the API. The MCP server has zod schemas on every tool, so partial credit (15 of 25). llms.txt at docs.brightdata.com with Markdown pages (10). MCP descriptions say what each tool returns, and the dataset tools say when to use them instead of run_scraper (14 of 20). Typed inputs with enums such as engine (bing or google), time_range and format (12 of 15). Per-product error catalogues for Web Unlocker, Browser API and scrapers, and curl examples (13 of 15). Release notes, now dated to 27 October 2026, and an MCP CHANGELOG with a gap between 2.6.0 and 2.10.0, per the 1 October reading (10 of 15).

Crawl4AI 55

No OpenAPI or similar contract for the cloud API, and the MCP tool schemas are only in the live tools/list, which needs a key. /v1/prices is machine-readable but covers prices, not calls (0). llms.txt of 304 lines written for agents (10). Each endpoint says what it is for and when to choose another, such as trying without a proxy first, using browser only when auto misses content and /scrape/jobs for slow lists (16 of 20). Parameters are listed with types, defaults, allowed values (format both, md or html; proxy none, isp or residential) and limits, in prose (10 of 15). Curl examples per endpoint and documented error strings per status with whether each is charged (14 of 15). No changelog for the cloud API, only dated notes such as /serp 'added 8 Oct 2026'. The library has a dated CHANGELOG (5 of 15). 0 + 10 + 16 + 10 + 14 + 5 = 55.

Bright Data 79

Since MCP 3.1.0 (15 October 2026) all 32 tools are on by default, with groups= and tools= on the hosted URL or GROUPS and TOOLS locally to narrow them (5 for more than 30, plus 10 for the filters, 15). Markdown, text, HTML or JSON-LD output from the scrape tools, content cut at 200,000 characters, batch tools of up to 5 searches or 5 URLs and filterable dataset search, but no general field selection (15 of 20). Error codes with a retry-or-fix classification, and since 4.3.0 the local server uses the hosted server's codes (rate_limited, invalid_input, quota_exceeded) (17 of 20). Since 4.1.0 every tool sets readOnlyHint, destructiveHint and openWorldHint, with browser click and type marked destructive. Billing counts successful requests, and there are no idempotency keys (17 of 20, as firecrawl-mcp and apify-mcp). One required secret and auto-created zones, and official Python (brightdata-sdk) and JavaScript SDKs in the docs index (15). 15+15+17+17+15 = 79.

Crawl4AI 67

16 MCP tools, six of them for billing and support (15 of 25). format md or html, parse opt-in blocks, batch limits and job results paged 500 at a time over HTTP and 50 over MCP (16 of 20). String error codes an agent can branch on (blocked, login-wall, no-answer, busy, no_credit, plan_cap, spend_cap), each with whether it is charged (18 of 20). Failed and cancelled calls aren't charged and jobs_submit takes an idempotency key, but no tool annotations are documented and two tools can charge the saved card (10 of 20). Only url or q is required. There is no official SDK for the cloud API; the docs give curl and standard-library Python and Node examples (8 of 15). 15 + 16 + 18 + 10 + 8 = 67.

Bright Data 61

The API is closed under the Bright Data licence, and the MCP server is MIT, so we gave 20 between the closed and open bands (20 of 30). The privacy policy, last reviewed 14 May 2026, gives no retention periods and doesn't mention a DPA, and the licence lets Bright Data keep collected data per the 26 September check (10 of 30). Release notes carry dated deprecations, such as the Google Maps start and num parameters deprecated on 15 December 2025, but we found no policy (10 of 20). No subprocessor list. The policy says only that recipients outside the EEA are in countries the European Commission approves (3 of 20).

Crawl4AI 70

The library is Apache-2.0 and the cloud is closed under terms that name CONTEXT4AI PTE LTD (22 of 30). The privacy policy gives retention periods (account plus 90 days, usage records 13 months, logs 90 days), but none for crawled content, and llms.txt says repeated URLs are served from a shared archive while the security page says fetched pages are used only to answer the request (18 of 30). The legacy cloud API closed on 6 November 2026 with a migration guide but no dated notice found, and the terms promise notice of material changes only 'where reasonable' (8 of 20). Subprocessors named with their locations on the security page, with EU storage of account records (18 of 20). 22 + 18 + 8 + 18 = 66.

Bright Data 40

No public status page found. status.brightdata.com doesn't resolve and the docs index and security overview link none (0). No readable incident history, so 5 by the rubric (5). The SERP API fair-use page gives 1,000 requests a minute for unfunded accounts with HTTP 429 above it and a header that reads the account's limit, paid Web Unlocker plans state unlimited concurrency, and the free tier is 5,000 requests a month (15). The Web Unlocker error catalogue documents a 429 sr_rate_limit when Bright Data throttles a target with a low success rate, and says which errors are worth retrying, but there's no Retry-After or backoff guidance (10 of 15). The pricing page names a Premium SLA for Enterprise and claims 99.99 per cent uptime, with no SLA terms published (0). GA (10). 0+5+15+10+0+10 = 40.

Crawl4AI 35

Graded on the hosted lines, for Crawl4AI Cloud, which the README and llms.txt now lead with and which an agent calls through its MCP server; the library would read on the local lines. No status page is linked, and status.crawl4ai.com fails TLS (0). No incident history readable (5). Per-plan limits are published with numbers in /v1/prices, from 10 requests a minute and 2 at a time on Free to 1,500 and 50 on Scale (15). 429 carries X-RateLimit headers and a message, 503 busy says to resend after a few seconds, and MCP job submissions take an idempotency key, but no Retry-After or backoff guidance (10 of 15). No SLA found in the terms or pricing (0). Prices and the free credit are a soft launch, /serp is beta and /answer experimental, while scrape, extract and search carry no beta label (5 of 10). 0 + 5 + 15 + 10 + 0 + 5 = 35.

Bright Data 40

No x402, MPP or L402. Issue #126 asks for x402 and is open (0). Web Unlocker and SERP at $1.50 per 1,000 successful requests, $1.30 on Scale, published without a login (20). 5,000 requests a month, 'No credit card required' (20). A person has to sign up and create a key in the control panel (0).

Crawl4AI 45

No x402, MPP or L402 (0). Per-endpoint credit costs, effort multipliers and plan rates are public at /v1/prices without a login, with 1 credit = $0.001 (20). 1,000 free credits a month, no card (20). A 7-day pass key comes from a click on api.crawl4ai.com without an account, which is still a person in a browser, and a permanent key needs email verification (5 of 20). 0 + 20 + 20 + 5 = 45.