v0.2.0 - Podcasts, Vimeo, and remote media

This release expands the ingestion layer with three new source adapters.\n\nHighlights:\n- Podcast RSS ingestion with episode metadata, Podcasting 2.0 transcript support, enclosure download, and Whisper fallback\n- Apple Podcasts episode resolution through the public catalog lookup\n- Vimeo oEmbed metadata, subtitle extraction, and local media fallback\n- Direct audio and video URL ingestion with streamed downloads\n- HLS .m3u8 resolution through yt-dlp\n- Additional AAC, FLAC, OGG, and Opus transcription support\n- 23 offline tests across Python 3.11 and 3.12\n\nThe same adapters are installed in the local Hermes runtime and the knowledge backend has been restarted.

v0.1.1 - English documentation

This patch release adds the complete English documentation set:\n\n- English README with installation, configuration, MCP, Docker, and security guidance\n- English WeChat Channels integration guide\n- English contribution, security, and third-party notices\n- Package version bumped to 0.1.1\n\nNo runtime behavior was changed.

v0.1.0 - Initial public release

Initial public release of Hermes Knowledge Ingestion.\n\nHighlights:\n- Plugin-based source adapter architecture\n- Web, X/Twitter, YouTube, WeChat article/video, PDF, image, and local-file ingestion\n- AI summarization, classification, tagging, and knowledge linking\n- Obsidian Markdown and SQLite persistence\n- Hermes-compatible MCP tool interface\n- Local-first macOS workflow for WeChat Channels\n\nUsers provide their own OpenAI-compatible or local model endpoint and credentials. No model API key is bundled.