Skip to content

Headless service for watching, parsing, storing, and serving AI coding sessions

License

Notifications You must be signed in to change notification settings

yologdev/yocore

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

25 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Yocore

Headless service for watching, parsing, storing, and serving AI coding sessions.

Yocore is the core engine behind Yolog - a platform that archives and visualizes AI pair programming sessions from Claude Code, Cursor, and other AI coding assistants.

Features

  • Session Watching: Automatically watches folders for new AI coding sessions
  • Multi-Parser Support: Parses Claude Code, OpenClaw, and other AI assistant formats
  • Full-Text Search: SQLite FTS5-powered search across all sessions and memories
  • Memory System: Extract and organize decisions, facts, preferences, and tasks
  • HTTP API: RESTful API for session replay, search, and memory management
  • MCP Server: Model Context Protocol integration for AI assistants
  • Lifeboat Pattern: Session context preservation across context compaction

Installation

npm (Recommended)

npm install -g @yologdev/core

Homebrew (macOS)

brew install yocore

Binary Download

Download the latest release from GitHub Releases.

Build from Source

git clone https://github.com/yologdev/yocore.git
cd yocore
cargo build --release

Usage

Start the Server

# Start with default config
yocore

# Start with custom config
yocore --config /path/to/config.toml

# Start in MCP server mode (for Claude Code integration)
yocore --mcp

Configuration

Create a config file at ~/.config/yocore/config.toml:

[server]
port = 19420
host = "127.0.0.1"
# api_key = "optional-secret"  # Enable for remote access

[database]
path = "~/.local/share/yocore/yocore.db"

# Watch multiple paths with different parsers
[[watch]]
path = "~/.claude/projects"
parser = "claude_code"

[[watch]]
path = "~/.openclaw/workspace"
parser = "openclaw"

[ai]
enabled = false
# provider = "anthropic"
# ANTHROPIC_API_KEY from env

[ai.features]
title_generation = false
skills_discovery = false
memory_extraction = false

MCP Tools

When running in MCP mode (yocore --mcp), the following tools are available:

Tool Description
yolog_search_memories Hybrid keyword + semantic search
yolog_get_project_context Get project overview with categorized memories
yolog_get_memories_by_type Filter memories by type (decision, fact, etc.)
yolog_get_memories_by_tag Filter memories by tag
yolog_get_recent_memories Get memories from recent sessions
yolog_get_session_context Get session state with lifeboat pattern
yolog_save_lifeboat Save session state before context compaction

Claude Code Integration

For the best experience with Claude Code, use yoskill - a skill pack that provides easy access to Yocore's memory system.

Setup

  1. Start Yocore in MCP mode:

    yocore --mcp
  2. Add to your Claude Code MCP config (~/.claude/claude_desktop_config.json):

    {
      "mcpServers": {
        "yolog": {
          "command": "yocore",
          "args": ["--mcp"]
        }
      }
    }
  3. Install yoskill for convenient slash commands:

    git clone https://github.com/yologdev/yoskill ~/.claude/skills/yoskill

Available Skills

Skill Description
/yo context Get session context + relevant memories
/yo project Get project-wide context
/yo search <query> Search memories by keyword
/yo search tag:<name> Filter by tag (e.g., tag:bug)

License

MIT License - see LICENSE for details.

Links

About

Headless service for watching, parsing, storing, and serving AI coding sessions

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Contributors 3

  •  
  •  
  •