Skip to content

v2.0.7 — Crawl4AI auth and safer video extraction

Choose a tag to compare

View all tags
@lfnovo lfnovo released this 02 Sep 14:10
· 14 commits to main since this release
v2.0.7
1be297b
This commit was created on GitHub.com and signed with GitHub’s verified signature.
GPG key ID: B5690EEEBB952194
Verified
Learn about vigilant mode.

Patch release with two fixes: Crawl4AI Docker mode works again with Crawl4AI >= 0.9.0, and video extraction no longer writes next to your source file. Validated with the canonical unit + integration suite (308 tests), the packaging gate on all three surfaces (library import, CLI, MCP handshake) from the built wheel, and all optional extras resolving. Media e2e (real STT calls) is unverified this release due to a provider billing issue on the maintainer's side; the video pipeline is covered by unit tests.

✨ New features

Library / CLI / MCP

  • crawl4ai_api_token config setting, with the CRAWL4AI_API_TOKEN env var taking precedence, mirroring CRAWL4AI_API_URL (#80). Set it to authenticate against a Crawl4AI Docker instance.
  • CLI: content-core config list and config set now mask credential-like values (keys containing token, key, secret, password) so a stored token never lands on stdout (#80).

🐛 Notable fixes

  • Crawl4AI Docker mode was rejected by any Crawl4AI >= 0.9.0 instance (#80). Recent versions require a bearer token for external connections by default, and the client sent no Authorization header. When the token is configured it is sent as Authorization: Bearer <token>; unset keeps the previous unauthenticated behavior for older instances.
  • Video extraction wrote an intermediate <name>_audio.mp3 next to your source video and deleted it afterwards (#74). A pre-existing file with that name was silently overwritten and removed, read-only source directories failed, and concurrent extractions of the same video collided. The intermediate now lives in a temporary directory, like audio transcription already did.

⚠️ Breaking changes & deprecations

None.

🙏 Thanks

Thanks to the Open Notebook users who reported the Crawl4AI authentication failure downstream (lfnovo/open-notebook#1243).

Full details in the CHANGELOG.

Install: pip install content-core==2.0.7
Extras: content-core[docling], [crawl4ai], [langchain]
Zero-install: uvx content-core extract <source>
MCP: uvx content-core mcp

Assets 2
Loading