mj-audio-topics

You can automate audio transcription and topic extraction by using a Skill that processes Chinese investment audio files into structured Lark Wiki documents. It transcribes audio, segments topics using a 5-dimension scoring model, and appends summaries to relevant topic documents automatically.

The best way to organize extracted audio topics into a Lark Wiki knowledge base is through automated multi-tagging. The system appends a single topic to multiple relevant documents, tracks topic registries, and maintains atomic state files to support resumable batch processing.

Deep search verification for high-signal investment topics works through a two-phase workflow. Phase 1 transcribes and scores topics, while Phase 2 performs deep web search and comprehensive analysis on high-score topics selected by the user.

Yes, you need lark-cli to transcribe Chinese audio files and manage Wiki API access. The Skill requires lark-cli specifically for Lark Minutes transcription and Wiki API integration to build the structured topic documents.

Yes, you can resume batch processing if your audio transcription pipeline is interrupted. The system implements atomic state persistence that tracks processed files and topic registries, ensuring batch processing can resume without data loss.

To run the mj-audio-topics Skill for investment research, you need jq and lark-cli installed. You must also place your Chinese investment audio files in the designated input directory structure before invoking the processing command.