Skip to content

docs(api): document MIME and PDF processor options - #93

Merged
VinciGit00 merged 8 commits into
mainfrom
marcovinciguerra/docs/search-content-types
Jul 30, 2026
Merged

docs(api): document MIME and PDF processor options#93
VinciGit00 merged 8 commits into
mainfrom
marcovinciguerra/docs/search-content-types

Conversation

@VinciGit00

@VinciGit00 VinciGit00 commented Jul 29, 2026

Copy link
Copy Markdown
Member

Documents the current MIME allowlist and PDF processor contract across Scrape, Extract, Search, and Crawl.

Changes

  • replace obsolete contentTypes / content_types references with allowedTypes / allowed_types
  • explain that omitting the allowlist accepts every supported type; "all" and "*" are invalid
  • document independent PDF processors, the 25-page default, 1..500, and -1 unlimited
  • update API references, service guides, SDK tables, and cURL/Python/JavaScript snippets

Verification

  • git diff --check
  • generated snippets tested for four endpoints and three languages
  • cURL, Python SDK, and JavaScript SDK live local executions all succeeded for Scrape, Extract, Search, and Crawl

SDK counterparts: ScrapeGraphAI/scrapegraph-py#101 and ScrapeGraphAI/scrapegraph-js#21. Playground: ScrapeGraphAI/sgai-stack#963.

VinciGit00 and others added 4 commits July 28, 2026 15:53
Verified every claim against the live server at mcp.scrapegraphai.com and
the apps/mcp implementation.

- remove the "Star us on GitHub" card
- drop the local stdio path: `pip install scrapegraph-mcp` resolves to PyPI
  1.0.1, which exposes 8 pre-v3 tools (markdownify, smartscraper, ...) and
  does not match the v2 API this section documents
- drop the Smithery page for the same reason, with a redirect to the
  introduction and inbound links repointed
- document the transport honestly: Streamable HTTP only, SSE unsupported,
  60s per tool call
- document all three accepted API key headers, plus OAuth 2.1 with PKCE
  S256 and the mcp:use scope
- correct the tool table to /api/* paths with real defaults and limits
- replace the migration note's dead v3.0.0 release link with a legacy
  name mapping table
- collapse the four client pages onto one shape: configure, authenticate,
  API key alternative, verify with `credits`
- bump the Claude Desktop mcp-remote pin from 0.1.25 to 0.1.38
Co-Authored-By: Claude Opus 5 (1M context) <noreply@anthropic.com>
@VinciGit00 VinciGit00 changed the title docs(search): document contentTypes filter docs(search): document MIME and PDF processor options Jul 30, 2026
@VinciGit00 VinciGit00 changed the title docs(search): document MIME and PDF processor options docs(api): document MIME and PDF processor options Jul 30, 2026
@VinciGit00

Copy link
Copy Markdown
Member Author

Update

Changed:

  • documented that PDF processing costs 1 credit per page actually processed
  • applied the wording consistently across Scrape, Extract, Search, Crawl, API reference, and both SDK guides

Test plan:

  • npx --yes mintlify@latest validate
  • local Mintlify preview: all 9 changed endpoint/SDK pages returned HTTP 200 and rendered the pricing copy
  • git diff --check
  • global broken-link check (29 pre-existing legacy links outside this delta)

@VinciGit00
VinciGit00 merged commit b996582 into main Jul 30, 2026
1 check passed
@VinciGit00
VinciGit00 deleted the marcovinciguerra/docs/search-content-types branch July 30, 2026 15:03
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant