← Back to Skills Marketplace
ukkit

Memcord

by Neeraj Tikku · GitHub ↗ · v4.0.1 · MIT-0
cross-platform ⚠ suspicious
50
Downloads
1
Stars
0
Active Installs
2
Versions
Install in OpenClaw
/install memcord
Description
Privacy-first, self-hosted chat memory for OpenClaw — save and recall conversation history across sessions without any cloud dependency.
README (SKILL.md)

Memcord — Persistent Chat Memory

Memcord is a self-hosted MCP server that gives you persistent, searchable memory across conversations. All data is stored as plain JSON files on your own machine — nothing leaves your device.

Adding Memcord to Your OpenClaw Config

Merge the following into your ~/.openclaw/openclaw.json. Only two tools are exposed: memcord_auto_save (write) and memcord_read (read).

// ~/.openclaw/openclaw.json
{
  mcp: {
    servers: {
      memcord: {
        command: "uvx",
        args: ["memcord"],
        toolFilter: {
          include: ["memcord_auto_save", "memcord_read"]
        }
      }
    }
  }
}

To use a custom slot name instead of the default "default" slot, add the env field:

// ~/.openclaw/openclaw.json
{
  mcp: {
    servers: {
      memcord: {
        command: "uvx",
        args: ["memcord"],
        env: { MEMCORD_DEFAULT_SLOT: "main" },
        toolFilter: {
          include: ["memcord_auto_save", "memcord_read"]
        }
      }
    }
  }
}

How to Use This Skill

When activated, this skill gives you two tools:

memcord_auto_save — Save the current conversation to persistent memory. Call this at the end of meaningful conversations or whenever the user asks you to remember something. No slot setup required.

memcord_read — Recall everything stored in memory. Call this at the start of a new conversation to resume context, or when the user asks "what do you remember?" or "what did we discuss before?".

When to auto-save

  • User says "remember this", "save this", "keep this in mind"
  • End of a session with important decisions or conclusions
  • After completing a task the user may reference again

When to read

  • User says "what do you remember?", "remind me", "what did we discuss?"
  • Start of a new session when continuity is expected
  • Before answering questions that likely depend on past context
Usage Guidance
Review before installing. This skill is not showing exfiltration or destructive behavior, but it can persist chat contents locally and recall them later. Use it only if you are comfortable with local conversation retention, avoid saving secrets or credentials, and prefer explicit 'remember this' use until the package documents storage location, deletion, retention, and consent behavior clearly.
Capability Assessment
Purpose & Capability
The stated purpose is persistent, searchable chat memory, and the exposed capabilities match that purpose: one write tool and one read tool for local memory.
Instruction Scope
The skill tells the agent to save the current conversation at the end of meaningful sessions and after tasks the user may reference again, which is broader than explicit user-directed requests like 'remember this'.
Install Mechanism
Installation is disclosed as a uv/uvx package install for the external package 'memcord'; the artifact contains only SKILL.md and no bundled executable code.
Credentials
The OpenClaw example limits the MCP tool surface to memcord_auto_save and memcord_read, and the artifact claims data stays on the user's machine, which is proportionate for a memory tool.
Persistence & Privilege
The skill discloses that all data is stored as plain local JSON, but it does not specify storage path, deletion/retention controls, first-save consent, or warnings to avoid secrets and sensitive personal data.
How to Use
  1. Make sure OpenClaw is installed (local or Docker)
  2. Run the install command in chat: /install memcord
  3. After installation, invoke the skill by name or use /memcord
  4. Provide required inputs per the skill's parameter spec and get structured output
Version History
v4.0.1
- Updated OpenClaw integration instructions for memcord, including new configuration structure. - The `command` is now set to `"uvx"` with `"args": ["memcord"]` in setup examples. - Removed explicit mention of installed binaries and cleaned up metadata requirements. - Documentation for custom slot usage was revised for clarity.
v4.0.0
Major release: memcord 4.0.0 introduces a privacy-first, self-hosted chat memory for OpenClaw. - Local, persistent chat history with no cloud dependency; all data saved as local JSON files. - Two tools exposed: memcord_auto_save (write) and memcord_read (read). - Easy gateway integration with support for custom memory slot names via environment variable. - Detailed usage guidance for when to save and recall memory. - Self-contained installation using the "uv" package manager.
Metadata
Slug memcord
Version 4.0.1
License MIT-0
All-time Installs 0
Active Installs 0
Total Versions 2
Frequently Asked Questions

What is Memcord?

Privacy-first, self-hosted chat memory for OpenClaw — save and recall conversation history across sessions without any cloud dependency. It is an AI Agent Skill for Claude Code / OpenClaw, with 50 downloads so far.

How do I install Memcord?

Run "/install memcord" in the OpenClaw or Claude Code chat to install it in one step — no extra setup required.

Is Memcord free?

Yes, Memcord is completely free, licensed under MIT-0. You can download, install and use it at no cost.

Which platforms does Memcord support?

Memcord is cross-platform and runs anywhere OpenClaw / Claude Code is available (cross-platform).

Who created Memcord?

It is built and maintained by Neeraj Tikku (@ukkit); the current version is v4.0.1.

💬 Comments