Memory LibSQLSAFE
🧠 High-performance persistent memory system for Model Context Protocol (MCP) powered by libSQL. Features vector search, semantic knowledge storage, and efficient relationship management - perfect for AI agents and knowledge graph applications.
Overview
From the repository's own README, as read at the audited commit. Badges and raw HTML are left out.
A high-performance, persistent memory system for the Model Context Protocol (MCP) powered by libSQL. This server provides vector search capabilities and efficient knowledge storage using libSQL as the backing store.
Features
- 🚀 High-performance vector search using libSQL
- 💾 Persistent storage of entities and relations
- 🔍 Semantic search capabilities
- 🔄 Knowledge graph management
- 🌐 Compatible with local and remote libSQL databases
- 🔒 Secure token-based authentication for remote databases
Configuration
This server is designed to be used as part of an MCP configuration. Here are examples for different environments:
Cline Configuration
Add this to your Cline MCP settings:
{
"mcpServers": {
"mcp-memory-libsql": {
"command": "npx",
"args": ["-y", "mcp-memory-libsql"],
"env": {
"LIBSQL_URL": "file:/path/to/your/database.db"
}
}
}
}Claude Desktop with WSL Configuration
For a detailed guide on setting up this server with Claude Desktop in WSL, see Getting MCP Server Working with Claude Desktop in WSL.
Add this to your Claude Desktop configuration for WSL environments:
{
"mcpServers": {
"mcp-memory-libsql": {
"command": "wsl.exe",
"args": [
"bash",
"-c",
"source ~/.nvm/nvm.sh && LIBSQL_URL=file:/path/to/database.db /home/username/.nvm/versions/node/v20.12.1/bin/npx mcp-memory-libsql"
]
}
}
}Database Configuration
The server supports both local SQLite and remote libSQL databases through the LIBSQL_URL environment variable:
For local SQLite databases:
{
"env": {
"LIBSQL_URL": "file:/path/to/database.db"
}
}For remote libSQL databases (e.g., Turso):
{
"env": {
"LIBSQL_URL7a66e1608ad7OBSERVED · 2026-10-09Connect
Built from this server's own package name, version and transport as found in its source — not copied from anyone's documentation, so it cannot drift against a page we do not control. Replace the environment placeholders with a token scoped to the least it needs.
claude mcp add mcp-memory-libsql --env DATABASE_AUTH_TOKEN=${DATABASE_AUTH_TOKEN} -- npx -y [email protected]{
"mcpServers": {
"mcp-memory-libsql": {
"command": "npx",
"args": [
"-y",
"[email protected]"
],
"env": {
"DATABASE_AUTH_TOKEN": "${DATABASE_AUTH_TOKEN}"
}
}
}
}Exposed tools (6)
2 read · 2 write · 2 destructive. Blast radius: 2 tools can delete or overwrite — an agent that can be talked into calling a tool can be talked into calling this one.
| Tool | Risk | Description |
|---|---|---|
create_entities | write | Create new entities with observations and optional embeddings |
create_relations | write | Create relations between entities |
delete_entity | destructive | Delete an entity and all its associated data (observations and relations) |
delete_relation | destructive | Delete a specific relation between entities |
read_graph | read | Get recent entities and their relations |
search_nodes | read | Search for entities and their relations using text or vector similarity |
Trust audit
SAFEgrade B · trust 89/100 Nothing in the source contradicts what it says it does. Grade A is reserved for packages that have also passed the behavioural sandbox.
| Layer | What it checks | Result |
|---|---|---|
| L0 | Provenance & inventory | PASS |
| L1 | Static analysis of the code | PASS |
| L2 | Instruction surface (what it tells the agent) | PASS |
| L3 | Class-specific surface | WARN |
| L4 | Behavioural (sandbox) | SKIPPED |
What the source does
- Filesystem
- declared (6 observation(s))
- Network
- none-observed
- Shell
- none-observed
- Dependencies
- not all pinned
- Secrets in source
- none-found
Findings (9)
delete_entity, delete_relation
.prettierignore
import { databaseConfig } from '../../config/index.js';import { logger } from '../../utils/logger.js';import { EMBEDDING_DIMENSION } from '../../services/embedding-service.js';import { databaseConfig } from '../../config/index.js';import { logger } from '../../utils/logger.js';@libsql/client, @types/cors, @types/express, @types/node, @types/uuid, @xenova/transformers, cors, dotenv
models/Xenova/bge-small-en-v1.5/onnx/model_quantized.onnx
Gates applied: no_behavioural_pass.
7a66e1608ad7full audit observations/trust-audit/mcp-server/joleyline__memory-libsql.json · Report an issue / request a re-scanAudit history
Every audit this server has had. A grade with a past is a grade somebody is still checking.
| Date | Source | Verdict | Grade | Score | Change |
|---|---|---|---|---|---|
| 2026-10-09 | 7a66e1608ad7 | SAFE | B | 89 | first audit |
Questions
What is the Memory LibSQL MCP server?
🧠 High-performance persistent memory system for Model Context Protocol (MCP) powered by libSQL. Features vector search, semantic knowledge storage, and efficient relationship management - perfect for AI agents and knowledge graph applications.
What tools does Memory LibSQL expose?
6 in total: 2 read-only, 2 that write, and 2 that can delete or overwrite (delete_entity, delete_relation). Every one is listed on this page with its risk.
Is Memory LibSQL safe to connect to an agent?
The audit found nothing in the source that contradicts what it says it does, and graded it B (89/100). Grade A is held back for packages that have also passed a sandboxed behavioural run, which is why a clean server reads B. Separately from the audit: 2 of its tools can destroy data, so scope the token you give it to what you actually need.
What credentials does Memory LibSQL need?
It reads DATABASE_AUTH_TOKEN from the environment. Give it a token scoped to the least it needs — an agent that can be talked into calling a tool can be talked into calling it with your credentials.
How does Memory LibSQL run?
It speaks stdio, so it runs as a local process your client starts. It is published on npm as mcp-memory-libsql at 0.0.14.
How current is this page?
The grade is for one exact copy of the source (7a66e1608ad7), read on 2026-10-09. The repository is watched and re-audited when it changes.