- Extract structured JSON from SvelteKit kit.start() data arrays
- Convert JS object literals (unquoted keys) to valid JSON
- Data appears in structured_data field (machine-readable)
- License changed from MIT to AGPL-3.0
- Bump to v0.3.4
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Migrated webclaw-fetch from webclaw-tls (patched rustls/h2/hyper/reqwest)
to wreq by @0x676e67. wreq uses BoringSSL for TLS and the http2 crate
for HTTP/2 fingerprinting — battle-tested with 60+ browser profiles.
This removes all 5 [patch.crates-io] entries that consumers previously
needed. Browser profiles (Chrome 145, Firefox 135, Safari 18, Edge 145)
are now built directly on wreq's Emulation API with correct TLS options,
HTTP/2 SETTINGS ordering, pseudo-header order, and header wire order.
84% pass rate across 1000 real sites. 384 unit tests green.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- --cookie-file reads Chrome extension format ([{name, value, domain, ...}])
- Works with EditThisCookie, Cookie-Editor, and similar browser extensions
- Merges with --cookie when both provided
- MCP scrape tool now accepts cookies parameter
- Closes#7
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Response.headers() now returns &http::HeaderMap instead of
&HashMap<String, String>. Updated FetchResult, is_pdf_content_type,
is_document_content_type, is_bot_protected, and all related tests.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
primp 1.2.0 moved to reqwest 0.13 and now patches reqwest itself
(primp-reqwest). Without this patch, cargo install gets vanilla
reqwest 0.13 which is missing the HTTP/2 impersonation methods.
Users should use: cargo install --locked --git ...
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
- Docker image auto-built on every release via CI
- QuickJS sandbox executes inline <script> tags to extract JS-embedded
content (window.__PRELOADED_STATE__, self.__next_f, etc.)
- Bumped version to 0.2.1
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
webclaw-core has QuickJS behind a feature flag for extracting data from
inline <script> tags (window.__PRELOADED_STATE__, self.__next_f, etc).
The server was using an old lockfile without the feature enabled.
Updated deps to v0.2.0 and explicitly enabled quickjs. This improves
extraction on SPAs like NYTimes, Nike, and Bloomberg where content is
embedded in JS variable assignments rather than visible DOM.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Document extraction:
- DOCX: auto-detected, outputs markdown with headings (via zip + quick-xml)
- XLSX/XLS: markdown tables with multi-sheet support (via calamine)
- CSV: quoted field handling, markdown table output
- All auto-detected by Content-Type header or URL extension
New features:
- -f html output format (sanitized HTML)
- Multi-URL watch: --urls-file + --watch monitors all URLs in parallel
- Batch + LLM: --extract-prompt/--extract-json works with multiple URLs
- Mixed batch: HTML pages + DOCX + XLSX + CSV in one command
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
--only-main-content, --include, and --exclude were ignored in batch
mode because run_batch used default ExtractionOptions. Added
fetch_and_extract_batch_with_options to pass CLI options through.
Closes#3
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Watch mode:
- --watch polls a URL at --watch-interval (default 5min)
- Reports diffs to stdout when content changes
- --on-change runs a command with diff JSON on stdin
- Ctrl+C stops cleanly
Webhooks:
- --webhook POSTs JSON on crawl/batch complete and watch changes
- Auto-detects Discord and Slack URLs, formats as embeds/blocks
- Also available via WEBCLAW_WEBHOOK_URL env var
- Non-blocking, errors logged to stderr
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Adds --output-dir flag for CLI. Each extracted page gets its own file
with filename derived from the URL path. Works with single URL, crawl,
and batch modes. CSV input supports custom filenames (url,filename).
Root URLs use hostname/index.ext to avoid collisions in batch mode.
Subdirectories created automatically from URL path structure.
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Embeds QuickJS (rquickjs) to execute inline <script> tags and extract
data hidden in JavaScript variable assignments. Captures window.__*
objects like __preloadedData (NYTimes), __PRELOADED_STATE__ (Wired),
and self.__next_f (Next.js RSC flight data).
Results:
- NYTimes: 1,552 → 4,162 words (+168%)
- Wired: 1,459 → 9,937 words (+580%)
- Zero measurable performance overhead (<15ms per page)
- Feature-gated: disable with --no-default-features for WASM
Smart text filtering rejects CSS, base64, file paths, code strings.
Only readable prose is appended under "## Additional Content".
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Crawl:
- Real-time progress on stderr as pages complete
- --crawl-state saves progress on Ctrl+C, resumes from saved state
- Visited set + remaining frontier persisted for accurate resume
MCP server:
- Reads WEBCLAW_PROXY and WEBCLAW_PROXY_FILE env vars
- Falls back to proxies.txt in CWD (existing behavior)
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
Critical:
- MCP server identifies as "webclaw-mcp" instead of "rmcp"
- Research tool poll loop capped at 200 iterations (~10 min)
CLI:
- Non-zero exit codes on errors
- Text format strips markdown table syntax
MCP server:
- URL validation on all tools
- 60s cloud API timeout, 30s local fetch timeout
- Diff cloud fallback computes actual diff
- Batch capped at 100 URLs, crawl at 500 pages
- Graceful startup failure instead of panic
Co-Authored-By: Claude Opus 4.6 (1M context) <noreply@anthropic.com>
