A specialized skill for AI agents to retrieve, search, and analyze voice recordings, transcripts, and AI-generated summaries from Pocket AI devices.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install heypocket-reader
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install heypocket-reader using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Pocket Transcripts?
Pocket Transcripts is a bridge between your physical Pocket AI recording hardware and your AI-driven workspace. By leveraging a reverse-engineered API, this skill allows users to programmatically access voice data stored on heypocket.com. As part of the Openclaw Skills ecosystem, it empowers AI agents to process raw recordings into actionable insights, making it an essential tool for those who rely on high-fidelity voice notes for their daily productivity.
Whether you need to pull a specific meeting transcript or search through months of voice memos, this skill provides the necessary endpoints to interface with Pocket's cloud infrastructure. By integrating with Openclaw Skills, you can automate the flow of voice data directly into your notes, project management tools, or knowledge bases.
Pocket Transcripts Use Cases
- Searching through historical voice recordings for specific keywords or topics discussed in meetings.
- Automatically extracting action items from recent Pocket AI recordings to populate task managers.
- Generating consolidated reports by analyzing multiple voice summaries from a single project phase.
- Archiving raw transcript text into personal knowledge management systems for permanent storage.
- Retrieving location-based voice notes to reconstruct travel logs or field research.
How Pocket Transcripts Works
- The agent initializes the connection using a locally stored Firebase Bearer token extracted from a logged-in browser session.
- When a user requests recording data, the skill queries the Pocket AI production API with specific filters like date ranges or search terms.
- The API returns a list of recordings or detailed metadata including transcription objects and summarization models.
- The skill parses the complex JSON response, extracting Markdown summaries and raw text for the agent to use.
- The AI agent processes this data to answer queries, summarize conversations, or trigger follow-up automations within the Openclaw Skills framework.
Pocket Transcripts Setup
To use this within the Openclaw Skills framework, follow these steps to authenticate your session:
- Launch Chrome with your user profile to enable session persistence:
~/.claude/skills/browser/start.js --profile
- Navigate to the Pocket web app and log in manually:
~/.claude/skills/browser/nav.js https://app.heypocket.com
- Run the extraction script to save your authentication token to
~/.pocket_token.json:
python3 scripts/reader.py extract
Note: Tokens are valid for approximately one hour and require a refresh using the extraction script when expired.
Pocket Transcripts Data Schema & Taxonomy
The skill utilizes the PocketRecording and PocketSummarization data models to organize retrieved information.
PocketRecording Attributes
| Field | Type | Description |
|---|---|---|
| id | String | Unique identifier for the recording |
| title | String | User-defined or auto-generated title |
| recorded_at | DateTime | Timestamp of the audio capture |
| duration_str | String | Human-readable length (e.g., '5:30') |
| tags | List | Array of strings for categorization |
Content Metadata
- Transcription: Raw text string found in the data transcription path.
- Summarization: Markdown-formatted text providing a structured overview of the recording.
- Action Items: A list of discrete tasks identified by the Pocket AI engine.
Pocket Transcripts Advanced Features
- Reverse-engineered API support for direct cloud access without relying on standard export features.
- Full support for multi-speaker transcription data and location metadata (latitude/longitude).
- Integrated search functionality allows for deep queries across thousands of voice recordings within the Openclaw Skills environment.
- Automated extraction of Firebase Bearer tokens from browser IndexedDB for seamless authentication workflows.
- High-fidelity Markdown summary retrieval compatible with modern documentation tools.
SKILL.md
Loading
Related Openclaw Skills
Dyson CLI
A command-line interface for managing Dyson air purifiers, fans, and heaters through local MQTT communication.
Whale Pulse Monitor
Real-time BTC Orderbook Imbalance tracking to detect whale activity using Sovereign-Futures logic.
Sovereign Liquidity Sniffer
An institutional-grade L2 Orderbook Imbalance detector for identifying market maker footprints across top exchanges.
OpenClaw MCP Debugger
A high-performance diagnostic suite designed to debug, monitor, and repair Model Context Protocol (MCP) connections for AI agents.
Obsidian Master Assistant
A comprehensive master-grade configuration skill for Obsidian that integrates professional knowledge management methodologies with advanced AI automation workflows.
Crabukit
Crabukit is a dedicated security auditing tool that performs static analysis on skills to prevent the installation of malicious or vulnerable code.
Featured*
