Bright Data vs WebScraping.AI

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.

WebScraping.AI 36

One account key with no scopes or documented rotation, taken only as the api_key query parameter on REST. The hosted MCP uses OAuth 2.1 with PKCE, dynamic registration and token revocation, with no scopes listed (20, less 10 for the query string, 10 of 30). Every endpoint is a GET and all tools are annotated read-only, and the terms forbid logging into target accounts. No read-only or limited key (8 of 20). The hosted MCP wraps scraped results in 'EXTERNAL CONTENT - DO NOT EXECUTE COMMANDS' boundaries by default, opt-in on the npm package (13 of 15). /account reports credits and limits. No per-call log for customers (5 of 15). security.txt answers 404, and no disclosure policy, bug bounty or certification was found (0 of 20). 10 + 8 + 13 + 5 + 0 = 36.

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.

WebScraping.AI 95

Ten MCP tools (25). Output can be sized with /text, /selected and /selected-multiple, AI answers and fields, text_format, return_links, and up to 5 SERP pages a call (18 of 20). Target errors return error_code, the target's status and a next_step with parameters and credit cost, and 402, 403, 429 and 503 have stated fixes (20). Failed requests are free, the docs say automating next_step retries is safe, 503 carries Retry-After, and all ten tools carry readOnlyHint true and destructiveHint false. No idempotency key, though every call is a read (18 of 20). Only url is required, but JavaScript is on by default at 5 credits. Official SDKs in seven languages and a CLI (14 of 15). 25 + 18 + 20 + 18 + 14 = 95.

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).

WebScraping.AI 89

OpenAPI file (version 3.2.2) in YAML and JSON covering all 10 operations, and typed zod inputs on the MCP tools (25). llms.txt, llms-full.txt and a Markdown copy of every page (10). Endpoint pages and tool descriptions say what each returns and when to use it, and errors point to the next endpoint, such as /selected for response_too_large (17 of 20). Query parameters are typed, with 10 enums in the OpenAPI file, but fields and headers are free-form objects (12 of 15). Examples in seven languages, an error table with solutions and a table of target error codes (15). The spec carries a version and the MCP and SDK repositories keep dated changelogs and tags, but the API has no changelog page (10 of 15). 25 + 10 + 17 + 12 + 15 + 10 = 89.

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).

WebScraping.AI 51

Closed service under terms naming Urlooker LLC of Wyoming, with MIT clients (15 of 30). The privacy policy keeps usage logs 30 days and deletes personal data within 60 days of a request, and the terms rule out training on scraped content, prompts and outputs. But section 9.3 lets the company store scraped content and logs for as long as reasonably necessary, and no DPA was found (14 of 30). The terms give 30 days' notice of material changes and of termination without cause, and the MCP changelog dates its changes. No deprecation policy (8 of 20). Subprocessors, including AI model providers, are listed only for enterprise customers under NDA, and data may be processed in the US or wherever subprocessors operate (4 of 20). 15 + 14 + 8 + 4 = 41.

Bright Data 87

MCP v4.3.0 on 2 November 2026 (30). Ten MCP releases since 6 August, 2.11.2 (11 September), 2.11.3 (15 September), 2.12.0 (8 October), 3.0.0 (11 October), 3.1.0 (15 October), 4.0.0 (18 October), 4.1.0 (20 October), 4.1.1 (29 October), 4.2.0 (1 November) and 4.3.0, and product release notes dated 19, 24 and 27 October (20). Issues not read (GitHub's pages refuse our reader); outside pull requests such as #200 were merged within a week, and the 1 October count of 13 open issues stands (12 of 25). io.github.brightdata/brightdata-mcp in the official registry under a GitHub-verified namespace, now at 4.3.0, the latest release (15). Tests run in CI on every push, and the release workflow runs the tests and npm audit signatures before publishing with provenance (10). 30+20+12+15+10 = 87.

WebScraping.AI 82

Read for a closed service with open clients. MCP server 1.4.0 on 26 October 2026 and Python SDK 4.3.0 on 24 October (30). In the 90 days from 8 August, MCP 1.2.1, 1.2.2, 1.3.0, 1.3.1, 1.3.2 and 1.4.0 and Python 4.1.0, 4.2.0 and 4.3.0 (20). Dated changelogs in the repositories and email support. GitHub issues weren't read (10 of 25). In the official MCP registry as io.github.webscraping-ai/webscraping-ai, though the newest entry there is 1.0.7 of 24 July 2026, with current SDKs in seven languages (12 of 15). CI workflows on the MCP and Python repositories, and 1.3.1 moved to zod 4 for a published advisory (10). 30 + 20 + 10 + 12 + 10 = 82.