heypocket-reader

The heypocket-reader skill provides users with programmatic access to recordings, transcripts, summaries, and action items from Pocket AI (heypocket.com) devices. It solves the challenge of efficiently retrieving and managing audio content from multiple recordings, allowing users to process and analyze their meetings or conversations without manually navigating the Pocket AI interface. By leveraging a reverse-engineered API, the skill enables direct access to both metadata and full recording content, streamlining workflows for professionals who rely on detailed audio documentation.

Key features of heypocket-reader include listing recent recordings, retrieving full recording data including transcripts, summaries, action items, and speaker profiles, and performing searches based on text queries or geographical location. While transcript search is not built-in due to the cost of multiple API calls, the skill provides a clear pattern for implementing custom search routines. The skill returns structured data objects like PocketRecording, containing attributes such as title, description, duration, timestamp, speaker count, location, and tags, which can be programmatically processed for reporting or analysis. It integrates seamlessly with Python, enabling scripting and automation for managing and querying recordings.

The skill is suitable for professionals, researchers, or teams who need to analyze meetings, interviews, or other audio sessions. Use cases include generating summaries of meetings, extracting action items for follow-up, monitoring discussions across multiple locations, and performing keyword searches across transcripts for research or compliance purposes. Target users are those who are comfortable with Python scripting and want to automate access to their Pocket AI recordings, such as project managers, data analysts, or knowledge workers who rely on efficient audio data management.