Mcp Memory Graph

Name: Mcp Memory Graph
Author: RetroRobAI

by RetroRobAIcommunityv0.1.1

A context-aware memory MCP server that stores, retrieves, and manages memories with authority weighting, conflict detection, and typed relationship edges. It uses local embeddings and SQLite for vector search and persistence, and can migrate from other memory services.

From the registry: Authority-weighted memory graph for AI agents with conflict detection and semantic search.

GitHub Repository

Install & configure

Prerequisites

required
Install the Python package
Install the MCP server from PyPI using pip.
```
$ pip install mcp-memory-graph
```

Install in your client

Paste into the AI you're chatting with — it will figure out the rest

Please install the `mcp-memory-graph` MCP server into my current AI client (that's you).

Required prerequisites (do these first if not already done):
- **Install the Python package** — Install the MCP server from PyPI using pip. Run: `pip install mcp-memory-graph`

Canonical MCP server config (stdio transport):
- command: `mcp-memory-graph`
- args: []
- optional environment variables:
  - `MEMORY_GRAPH_DB_PATH`: Path to the SQLite database file used to store memories. (example: `<path-to-memories.db>`)
  - `MEMORY_GRAPH_MODEL`: Sentence-transformers model name to use for embeddings. (example: `<model-name>`)
  - `MEMORY_GRAPH_DIM`: Embedding dimension size for the selected model. (example: `<embedding-dimension>`)
  - `MEMORY_GRAPH_CONFLICT_THRESHOLD`: Cosine similarity threshold above which memories are flagged as conflicting. (example: `<conflict-threshold>`)
  - `MEMORY_GRAPH_DEFAULT_RESULTS`: Default number of memories returned during retrieval. (example: `<default-result-count>`)

Note: Prefer the installed CLI entrypoint `mcp-memory-graph` over running `python server.py` directly. The README shows Claude Code config with only `MEMORY_GRAPH_DB_PATH`, but all configuration is provided via environment variables and defaults are documented. No API key is required because embeddings run locally.

Add this MCP server to my current client's config in the correct format for you. If you need secrets or credentials I haven't provided, ASK me — do not invent values or leave raw placeholders. After adding it, tell me how to verify the server is connected.

Note: Prefer the installed CLI entrypoint `mcp-memory-graph` over running `python server.py` directly. The README shows Claude Code config with only `MEMORY_GRAPH_DB_PATH`, but all configuration is provided via environment variables and defaults are documented. No API key is required because embeddings run locally.

Tools (8)

store_memory

Store a memory with conflict detection and optional auto-resolve

retrieve_memories

Search memories by semantic similarity weighted by authority

check_conflicts

Preview conflicts before storing a memory

update_memory

Update memory content or priority with supersession tracking

delete_memory

Soft delete a memory while preserving history

add_memory_edge

Manually add a typed relationship between memories

get_related_memories

Traverse the relationship graph for a memory

list_memories

List memories with filters such as status, type, and priority

Environment variables

MEMORY_GRAPH_DB_PATHrequiredSQLite database path for storing memories

Packages

mcp-memory-graph

pypiv0.1.1transport: stdio

Details

Language: Python
Released: May 4, 2026
Last commit: May 4, 2026
Registry name: io.github.RetroRobAI/mcp-memory-graph

Related Servers

Labelhead Artist Momentum

Trending hip-hop artist momentum scores across four cultural dimensions.

78.5k

PostHog MCP Server

Official PostHog MCP Server for product analytics, feature flags, experiments, and more.

37.7k

Openmetadata Mcp

Official OpenMetadata MCP: governed context and business semantics for AI assistants and agents.

14.9k