API Reference¶

Complete API specification for Gobbler MCP tools.

For the full detailed API specification with input schemas, output formats, and error scenarios, see API.md in the repository root.

Tool Summary¶

Content Conversion¶

Tool	Description
`transcribe_youtube`	Extract YouTube video transcript to markdown
`fetch_webpage`	Convert web page to markdown
`fetch_webpage_with_selector`	Extract specific content with CSS/XPath
`convert_document`	Convert PDF, DOCX, PPTX, XLSX to markdown
`transcribe_audio`	Transcribe audio/video files with Whisper

Browser Automation¶

Tool	Description
`browser_check_connection`	Check browser extension connection
`browser_list_tabs`	List tabs in Gobbler group
`browser_navigate_to_url`	Navigate to URL
`browser_execute_script`	Execute JavaScript in active tab
`browser_execute_script_in_tab`	Execute JavaScript in specific tab
`browser_extract_current_page`	Extract current page as markdown

Crawling & Sessions¶

Tool	Description
`create_crawl_session`	Create authenticated browser session
`crawl_site`	Recursively crawl website

YouTube¶

Tool	Description
`download_youtube_video`	Download video to local file

Batch Processing¶

Tool	Description
`batch_transcribe_youtube_playlist`	Process entire YouTube playlist
`batch_fetch_webpages`	Convert multiple web pages
`batch_transcribe_directory`	Transcribe all audio in directory
`batch_convert_documents`	Convert all documents in directory

Queue Management¶

Tool	Description
`get_job_status`	Check queued job status
`list_jobs`	List jobs in queue
`get_batch_progress`	Get batch operation progress

Common Patterns¶

Output Format¶

All tools return markdown with YAML frontmatter:

---
source: <URL or file path>
type: <youtube_transcript|webpage|document|audio_transcript>
word_count: <number>
converted_at: <ISO 8601 timestamp>
---

# Content Title

Markdown content...

Error Handling¶

Tools return structured errors:

{
  "isError": true,
  "content": [{
    "type": "text",
    "text": "Error message with actionable guidance"
  }]
}

Auto-Queue¶

Long-running operations support automatic background queuing:

transcribe_audio(file_path="/path/to/large.mp4", auto_queue=True)

Returns job ID for status checking:

get_job_status(job_id="abc123-def456")

Service Dependencies¶

Tool	Requires
`transcribe_youtube`	None (uses YouTube API)
`transcribe_audio`	None (local Whisper)
`fetch_webpage`	Crawl4AI Docker
`convert_document`	Docling Docker
`browser_*`	Browser extension
Queue tools	Redis