meridian-mcp

Featured

Build, configure, and debug the Meridian MCP server. Exposes app session data to AI tools via the Model Context Protocol.

AI & Automation 336 stars 23 forks Updated 5 days ago MIT

Install

View on GitHub

Quality Score: 91/100

Stars 20%
84
Recency 20%
100
Frontmatter 20%
70
Documentation 15%
100
Issue Health 10%
50
License 10%
100
Description 5%
100

Skill Content

# Meridian MCP Server Skill ## What It Is A TypeScript MCP server (`packages/meridian-mcp/`) that exposes Meridian's structured session database to any MCP-compatible AI tool (Claude Code, Claude Desktop, Cursor, etc.). It opens `~/.meridian/meridian.db` read-only using **sql.js** (pure WebAssembly SQLite — no native Node.js modules, works with any Node.js version) and provides tools for querying app sessions, focus time, and activity history. ## Build & Run ```bash cd packages/meridian-mcp # Install dependencies npm install # Build TypeScript → dist/ npm run build # Start server (stdio transport) node dist/index.js ``` **Prerequisite**: The Meridian daemon must be running and have produced at least one session. The MCP server returns an error message on tool calls (does not crash) if the DB is missing. ## Available Tools | Tool | Description | |------|-------------| | `get-sessions` | List completed app sessions for a date (default: today) | | `get-timeline` | Full day timeline including idle and sleep gaps | | `get-stats` | Daily productivity stats — focus/idle/away time, top apps | | `get-active-session` | Currently in-progress session (if daemon is running) | | `get-apps` | All-time app usage stats | | `search-sessions` | Search sessions by window title, OCR text, or audio | | `get-session-detail` | Full OCR, elements, and signals for a session ID (audio excluded — stored in DB but not sent to LLMs) | | `health-check` | ETL run status, cursor position, and tota...

Details

Author
Meridiona
Repository
Meridiona/meridian
Created
4 months ago
Last Updated
5 days ago
Language
Rust
License
MIT

Integrates with

Similar Skills

Semantically similar based on skill content — not just same category