Programmatic, CLI, and MCP access to Wispr Flow data
概览
WisprMCP provides programmatic access to your Wispr Flow local database, enabling you to query transcripts, search conversations, analyze usage patterns, and integrate with AI workflows through the Model Context Protocol (MCP). - : Added collect command for word extraction by day (similar to GranolaMCP) - : Added sane defaults - wispr list now defaults to last 2 days - : Fixed timestamp parsing for proper date filtering and time display - : Initial release with full database access, CLI tools, and MCP server - Zero external dependencies - built entirely with Python standard library - Comprehensive search and filtering capabilities - Rich CLI with colored output and multiple formats - JSON-RPC 2.
README
以下为原文 README 节选，完整内容请查看原文。
WisprMCP
A Python library for accessing Wispr Flow’s SQLite database with CLI and MCP server support.
WisprMCP provides programmatic access to your Wispr Flow local database, enabling you to query transcripts, search conversations, analyze usage patterns, and integrate with AI workflows through the Model Context Protocol (MCP).
Recent Updates
- 2025-07-04: Added collect command for word extraction by day (similar to GranolaMCP)
- 2025-07-04: Added sane defaults -
wispr listnow defaults to last 2 days - 2025-07-04: Fixed timestamp parsing for proper date filtering and time display
- 2025-07-04: Initial release with full database access, CLI tools, and MCP server
- Zero external dependencies - built entirely with Python standard library
- Comprehensive search and filtering capabilities
- Rich CLI with colored output and multiple formats
- JSON-RPC 2.0 compliant MCP server for AI integration
Complete Feature Overview
📚 Library Access
- Direct SQLite database access to Wispr Flow transcripts
- Rich data models for transcripts, conversations, and metadata
- Comprehensive search and filtering capabilities
- Zero external dependencies - pure Python standard library
🖥️ Command Line Interface (7 Commands)
| Command | Purpose | Key Features |
|---|---|---|
list |
List recent transcripts | Defaults to 2 days, rich table output, filtering |
show |
Show specific transcript | Full details, context, all text versions |
search |
Search transcript text | Highlighting, context, date filtering |
stats |
Database statistics | Overall stats, app breakdown, quality metrics |
apps |
App usage analysis | Usage by app, sorting, filtering |
export |
Export data | JSON/CSV/Markdown, conversation grouping |
collect |
Extract YOUR words | Daily files (YYYY-MM-DD.txt) for analysis |
Output Features:
- Rich colored output with semantic highlighting
- Multiple formats: table, JSON, text, markdown
- Advanced filtering: date, app, status, conversation, word count
- Sane defaults:
wispr listshows last 2 days automatically
🗣️ Word Collection (Like GranolaMCP)
- Extract YOUR spoken words into daily text files
- 4 output formats:
raw(pure text),words,sentences,entries - Smart filtering: Exclude apps, minimum word counts, date ranges
- Ready for analysis: Word clouds, linguistic analysis, voice training
- Example:
wispr collect ./my-words→ creates2025-07-04.txtwith YOUR words
🤖 MCP Server Integration (10 Tools)
| Tool | Purpose |
|---|---|
get_recent_transcripts |
Get most recent entries |
list_transcripts |
List with date/app filters |
search_transcripts |
Search for specific text |
get_transcript |
Get specific transcript by ID |
get_conversations |
Get grouped conversations |
get_conversation |
Get specific conversation |
安装
This server does not publish a one-line install command.
Open the repository installation guide 新手？先看 mcpskill.org 教程 →配置
{
"mcpServers": {
"WisprMCP": {
"command": "python",
"args": ["/path/to/WisprMCP/wispr_mcp/mcp/server.py"],
"cwd": "/path/to/WisprMCP"
}
}
}