Web data infrastructure
Collect once. Keep the evidence.
Nineteen local MCP tools cover search, scraping, crawling, mapping, document parsing, bounded browser interactions, durable jobs, artifacts and change monitoring.
Discovery
search_web
Discover public-web sources with keyless search, ranked links and provenance.
map_site
Discover URLs with scope controls, sitemap coverage and deduplication.
list_extraction_presets
Inspect supported field-extraction presets and their requirements.
Collection
scrape_page
Collect page content, structured fields, metadata and evidence.
scrape_batch
Collect many URLs with bounded concurrency and per-item status.
crawl_site
Run a bounded durable crawl with pages, artifacts and progress.
parse_document
Extract information from supported PDF, DOCX, XLSX, HTML and text documents.
browser_session
Create, inspect and close isolated local browser sessions.
browser_action
Navigate and interact with an authorized browser session.
Operations
get_job
Inspect durable job state, errors and progress.
cancel_job
Request safe cancellation of queued or running work.
get_crawl
Page through captured crawl results and completeness details.
get_artifact
Retrieve bounded content and integrity metadata by opaque artifact ID.
Monitoring
create_watch
Create a saved target for repeatable change detection.
check_watch
Run a requested check against an existing monitor.
set_watch_schedule
Opt into or pause a monitor's recurring checks.
list_watches
Browse saved monitors and their status.
compare_crawls
Compare snapshots and identify supported change evidence.
get_change_report
Inspect change reports with bounded pagination.