Skip to content
    ↑↓ select↵ openesc close
    English中文
    XingP14

    Woclaw

    v0.1.0Memory & Context
    opencode-woclaw

    WoClaw Hub plugin for OpenCode — shared memory, topic messaging, and multi-agent context sharing

    Actively maintained
    #opencode#opencode-plugin#woclaw#openclaw#multi-agent#shared-memory
    GitHub reponpm pageHomepage
    GitHub stars

    4

    Monthly installs

    5

    2 in 7 days

    Composite scoreSCORE

    30.3

    Multi-signal model

    Last commit

    9 days ago

    2026-08-10

    Install and configure

    opencode.json

    Writes to this project's opencode.json — applies to this repository only.

    opencode.json

    {
  "$schema": "https://opencode.ai/config.json",
  "plugin": ["opencode-woclaw@0.1.0"]
}

    opencode loads npm dependencies through its embedded runtime on startup and caches them locally — no manual global install needed.

    README.md
    View original→

    Shared memory and messaging hub for AI agents across OpenClaw, Claude Code, Gemini CLI, OpenAI Codex CLI, OpenCode, and Hermes Agent (roadmap).

    MIT License GitHub Stars CI Docker Hub Node.js npm (scoped) npm npm npm npm npm

    WoClaw is organized into three layers:

    • The Hub layer handles memory, topics, graph memory, auth, and storage.
    • The adapter layer connects OpenClaw, Claude Code, Gemini CLI, Codex, OpenCode, and tracks Hermes Agent support on the roadmap.
    • The documentation and site layer covers GitHub Pages, the dashboard, the inspector, and migration guides.

    By default, the Hub uses local SQLite storage. MySQL is optional.

    Core Capabilities

    • Shared memory with tags, TTL, version history, and keyword search
    • Topic pub/sub with @agent mentions, history, private topics, and federation
    • Multi-framework adapters for OpenClaw, Claude Code, Gemini CLI, OpenAI Codex CLI, OpenCode, and Hermes Agent (roadmap)
    • Hooks and migration tools for session start/stop/PreCompact and historical imports
    • GitHub Pages-friendly landing page, dashboard, and inspector
    • Graph Memory with temporal, entity, causal, and semantic relationships

    Quick Start

    1. Start the Hub

    docker pull xingp14/woclaw-hub:latest
docker run -d \
  --name woclaw-hub \
  -p 8082:8082 -p 8083:8083 \
  -e AUTH_TOKEN=change-me \
  --restart unless-stopped \
  xingp14/woclaw-hub:latest

    The default data file is /data/woclaw.sqlite. To switch to MySQL, set DB_TYPE=mysql and the corresponding MYSQL_* variables.

    2. Connect Your Agents

    • OpenClaw plugin: npm install xingp14-woclaw
    • Claude Code / Gemini CLI / OpenCode: use woclaw-hooks
    • OpenAI Codex CLI: use woclaw-codex
    • MCP clients: use woclaw-mcp
    • VS Code / Cursor: build the WoClaw extension locally from source (Marketplace publishing pending vsce publish from maintainer, publisher: XingP14)

    Hermes Agent support is on the roadmap; current releases focus on the OpenClaw plugin and hook-based integrations.

    Topic messages can target one or more agents with @agent syntax, for example woclaw send my-topic "@alice @bob review this".

    Detailed setup and configuration live in docs/README.md.

    3. Open the Web UI

    • GitHub Pages: https://xingp14.github.io/woclaw/
    • Local Hub dashboard: http://your-hub:8084

    Hermes Agent Support

    Hermes Agent integration is planned for v0.6+. See ROADMAP.md for details.

    What's planned:

    • Skills / shared-skills / workspace-agents / model-config migration
    • Messaging settings and memory compatibility assessment
    • Hook scripts for session lifecycle (start/stop)
    • Documentation and install guide

    Current status: Design and documentation in progress. Active development begins after v0.6 stable release.

    Documentation

    Links