Integrates local-first semantic memory, embeddings, and reranking models into OpenClaw as a secure MCP server.
Install & Download
1. ClawHub CLI
The fastest way to install a skill directly from the registry.
npx clawhub@latest install memkeeper-mcp-setup
2. Manual Installation
Copy the skill folder to one of these locations
~/.openclaw/skills/ <project>/skills/ Priority: Workspace > Local > Bundled
3. Prompt Installation
Copy this prompt to OpenClaw to install it automatically.
Help me install memkeeper-mcp-setup using Clawhub. If Clawhub is not installed, install it first (npm i -g clawhub).
Prefer to download?
Get the raw skill files in a ZIP archive.
What is Memkeeper Memory for OpenClaw?
Memkeeper is a local-first semantic memory integration designed for OpenClaw. It configures a local Model Context Protocol (MCP) server that utilizes local embedding and reranking models, keeping your sensitive data private and on-device by default. This integration is part of the broader ecosystem of Openclaw Skills that enable developer-centric, secure workflows.
By leveraging Memkeeper, developers can provide their AI agents with semantic memory capabilities. The system relies on local SQLite storage and specialized semantic models, avoiding dependencies on external cloud providers or API keys unless explicitly configured.
Memkeeper Memory for OpenClaw Use Cases
- Establishing a secure, local-first vector memory store for agent workflows.
- Reducing external API dependency and latency by using local mxBai embedding and reranking models.
- Debugging and diagnosing failed MCP connections using automated verification probes.
- Restricting destructive database operations like deletion or direct graph mutation to safe, operator-controlled steps.
How Memkeeper Memory for OpenClaw Works
- Local Model Setup: The integration downloads and initiates local embedding and reranking models, creating an isolated SQLite store.
- Sanity Check: Verification utilities run to check that semantic models are operational and fully initialized.
- MCP Registration: The Memkeeper binary is registered as an MCP server with OpenClaw, exposing a scoped 12-tool profile.
- Connection Probing: The connection is verified using built-in doctor utilities to ensure communication over stdio.
- State Management: The agent interacts with the memory store through defined APIs, and the store remains local unless migration is explicitly requested.
Memkeeper Memory for OpenClaw Setup
To set up Memkeeper on your local machine, run the following commands to install, configure, and register the service.
First, define variables and install the binary if it does not already exist:
MEMKEEPER_BIN="${MEMKEEPER_BIN:-$HOME/.local/bin/memkeeper}"
MEMKEEPER_STORE="${MEMKEEPER_STORE:-$HOME/.memkeeper/store.sqlite}"
if [ ! -f "$MEMKEEPER_BIN" ]; then
curl -fsSL https://raw.githubusercontent.com/teflon07/memkeeper/main/install.sh \
-o /tmp/memkeeper-install.sh
bash /tmp/memkeeper-install.sh
fi
Next, download the local models and initialize the semantic store:
"$MEMKEEPER_BIN" pull-models
"$MEMKEEPER_BIN" init --store "$MEMKEEPER_STORE" --json
MEMKEEPER_REQUIRE_SEMANTIC=1 \
"$MEMKEEPER_BIN" doctor --store "$MEMKEEPER_STORE" --json
Register the MCP server with a restricted tool profile using Openclaw Skills management:
openclaw mcp add memkeeper \
--command "$MEMKEEPER_BIN" \
--arg mcp \
--env "MEMKEEPER_STORE=$MEMKEEPER_STORE" \
--env "MEMKEEPER_REQUIRE_SEMANTIC=1" \
--include "stats,search,get,memory_list,entity_search,graph_neighbors,graph_context,dream_graph,remember,pack,candidate_submit,candidate_list" \
--timeout 30 \
--connect-timeout 30
Finally, verify the configuration:
openclaw mcp doctor memkeeper --probe --json
Memkeeper Memory for OpenClaw Data Schema & Taxonomy
The database structure and default environmental controls are defined as follows:
Database Storage
- Default Store Path:
$HOME/.memkeeper/store.sqlite - Database Engine: SQLite (Local-first, single-file)
- Embeddings Backend: mxBai (default, local)
Exposed Tool Profile (12 Tools)
| Tool Name | Purpose |
|---|---|
stats |
View memory statistics and store metrics |
search |
Perform standard semantic search over stored records |
get |
Retrieve specific memory entries by identifier |
memory_list |
List existing memory blocks |
entity_search |
Search for parsed semantic entities |
graph_neighbors |
Fetch related graph nodes |
graph_context |
Contextualize queries within the semantic graph |
dream_graph |
Consolidate and optimize graph links |
remember |
Write a new memory block to the store |
pack |
Compress and organize memory representations |
candidate_submit |
Propose potential items for retention |
candidate_list |
Inspect pending storage proposals |
Memkeeper Memory for OpenClaw Advanced Features
- Local-First mxBai Models: Uses embedded semantic retrieval and reranking engines on-device without leaking data.
- Scoped MCP Security: Disables functions such as
forgetand database mutation (relationship_upsert) by default, requiring operator opt-in. - Automated Diagnostic Probes: Built-in verification utilizing the Openclaw Skills interface to confirm connection status and model health without exposing sensitive keys.
SKILL.md
Loading
Related Openclaw Skills
OrcaTrace
OrcaTrace provides real-time Polymarket data feed intelligence, smart-money wallet calibration, and automated market research to power autonomous AI agents.
Hong Kong Insurance Product Data API (HK Insure MCP)
An MCP-compliant API providing structured data for over 260 Hong Kong insurance products, including internal rates of return (IRR) and historical dividend fulfillment statistics.
Log Parser to Metrics
A highly efficient utility to convert raw log files into Prometheus metrics, JSON, or CSV formats using customizable regex patterns.
Algorithm Filing Assistant
Openclaw Skills turns AI product details into the right CAC filing path, required materials, submission steps, and ongoing compliance actions.
Signalis
Signalis provides real-time global intelligence and AI-business narrative tracking for autonomous agents using pay-per-call blockchain payments.
Workflow DAG Validator
Openclaw Skills Workflow DAG Validator finds DAG cycles, missing dependencies, and parameter mismatches before bad workflows ship.
Featured*
