AI Memory

Pre-tested Argus MCP backend configurations for ai memory servers.

Memory Bank MCP (containerized)

View on GitHub

Structured memory bank for AI assistants. Runs in a container with outbound network access for API synchronization.

name: "Memory Bank MCP (containerized)"
description: "Structured memory bank for AI assistants. Runs in a container with outbound network access for API synchronization."

memory-bank-mcp:
  type: stdio
  command: npx
  args:
    - "-y"
    - "@diazstg/memory-bank-mcp"
    - "--mode"
    - "code"
    - "--username"
    - "${MEMORY_BANK_USERNAME}"
  container:
    enabled: true
    network: bridge

Memory Bank MCP (subprocess)

View on GitHub

Structured memory bank for code-mode AI assistants. Persists project context, decisions, and progress across sessions.

name: "Memory Bank MCP (subprocess)"
description: "Structured memory bank for code-mode AI assistants. Persists project context, decisions, and progress across sessions."

memory-bank-mcp:
  type: stdio
  command: npx
  args:
    - "-y"
    - "@diazstg/memory-bank-mcp"
    - "--mode"
    - "code"
    - "--username"
    - "${MEMORY_BANK_USERNAME}"

Memory Server (containerized — network isolated)

View on GitHub

Persistent knowledge graph. Runs in a fully network-isolated container — all storage is in-process, no external calls.

name: "Memory Server (containerized — network isolated)"
description: "Persistent knowledge graph. Runs in a fully network-isolated container — all storage is in-process, no external calls."

memory:
  type: stdio
  command: npx
  args:
    - "-y"
    - "@modelcontextprotocol/server-memory"
  container:
    enabled: true
    network: none

Memory Server (subprocess)

View on GitHub

Persistent in-process knowledge graph for storing and retrieving entities and relations across conversations.

name: "Memory Server (subprocess)"
description: "Persistent in-process knowledge graph for storing and retrieving entities and relations across conversations."

memory:
  type: stdio
  command: npx
  args:
    - "-y"
    - "@modelcontextprotocol/server-memory"

Qdrant Vector Memory (containerized)

View on GitHub

Semantic vector search memory for AI assistants using Qdrant. Store and retrieve memories by meaning rather than exact text. Runs in a container with outbound network access to reach your Qdrant instance.

name: "Qdrant Vector Memory (containerized)"
description: "Semantic vector search memory for AI assistants using Qdrant. Store and retrieve memories by meaning rather than exact text. Runs in a container with outbound network access to reach your Qdrant instance."
# Secrets: QDRANT_URL — set with: argus-mcp secrets set QDRANT_URL http://localhost:6333
# Optional: QDRANT_API_KEY for Qdrant Cloud — argus-mcp secrets set QDRANT_API_KEY <key>
# COLLECTION_NAME — set with: argus-mcp secrets set COLLECTION_NAME mcp-memory

qdrant-memory-container:
  type: stdio
  command: uvx
  args:
    - "mcp-server-qdrant"
  env:
    QDRANT_URL: "${QDRANT_URL}"
    COLLECTION_NAME: "${COLLECTION_NAME}"
  container:
    enabled: true
    network: bridge

Want to add a configuration? See the contributing guide.