v1.3.0

notebooklm-skill 1.3.0

This release rebuilds notebooklm-skill around the current notebooklm-py 0.7.x API and gives the CLI, pipelines, and MCP server one shared compatibility layer.

Highlights

• Five consistent CLI entry points with JSON-first results and meaningful exit codes
• Profile-aware authentication, including local Google Chrome/Edge login and safe profile logout
• Mixed URL, text, and file ingestion with bounded concurrency and per-source outcomes
• Complete fast/deep research lifecycle with polling, terminal-state checks, and bounded source import
• Generation, listing, and precise download support for 11 canonical artifact types
• Five real async pipelines and a 13-tool FastMCP server compatible with FastMCP 2.x and 3.x
• PEP 668-safe isolated installer and a standard, atomic Claude Code Skill installer
• Safer deletion, downloads, shell execution, RSS reads, trend integration, and loopback-only MCP HTTP
• Rebuilt bilingual documentation, examples, CI, packaging, and Trusted Publishing release automation

Verified

• 87 tests passing with 93.70% coverage
• Python 3.10, 3.11, 3.12, and 3.14 coverage across local and CI validation
• Ruff, mypy, ShellCheck, actionlint, CodeQL, wheel/sdist, Twine, and clean-install checks
• Live Google Chrome authentication and NotebookLM create/source/chat/citation/summary/research/artifact/download flows
• Live FastMCP tool invocation and research-to-social pipeline execution
• Fresh public PyPI install with all five commands, 13 MCP tools, and packaged Skill present

Upgrade

python -m pip install --upgrade notebooklm-skill==1.3.0

See the full changelog or compare v1.2.2...v1.3.0.