Memory Nexus Mcp
Give your agent durable semantic memory and knowledge-graph recall across sessions instead of re-explaining context every time you open the repo.
Overview
Memory Nexus MCP is an MCP server for the Build phase that provides persistent semantic memory, knowledge graphs, and Intelligence Packs for AI agents.
What is this MCP server?
- Persistent semantic memory for AI agents with knowledge graph backing
- Intelligence Packs extend packaged memory capabilities per catalog description
- npm package @memory-nexus/mcp v1.0.0 with stdio transport
- Requires NEXUS_API_KEY environment variable for Memory Nexus API access
- npm package @memory-nexus/mcp version 1.0.0
- stdio MCP transport
- NEXUS_API_KEY documented as required environment variable
What problem does it solve?
Agents lose thread-local context between sessions, so you repeat explanations and decisions instead of building on structured memory.
Who is it for?
Indie builders running multi-session agent workflows who need hosted semantic memory rather than ad-hoc markdown notes.
Skip if: One-off scripts, strictly offline-only setups without a Nexus API key, or teams that already centralize memory in their own vector DB with no MCP layer.
What do I get? / Deliverables
After you set NEXUS_API_KEY and register the server, your agent can store and retrieve semantic memory and graph-linked knowledge over MCP.
- Cross-session semantic memory reads and writes from the agent
- Knowledge-graph-backed recall via MCP tools
Recommended MCP Servers
Journey fit
Long-horizon agent memory is core Build infrastructure when you are composing agents that must remember decisions, entities, and prior work. Agent-tooling is the shelf for MCP memory layers with knowledge graphs and packaged Intelligence Packs, distinct from generic backend APIs.
How it compares
Hosted agent memory and knowledge-graph MCP, not local Taskwarrior or Python repo scaffolds.
Common Questions / FAQ
Who is Memory Nexus MCP for?
Solo builders and small teams building agent products who want durable semantic memory and knowledge graphs accessible from Claude Code, Cursor, or Codex.
When should I use Memory Nexus MCP?
Use it in Build when you are adding agent-tooling and need cross-session recall, structured entities, or Intelligence Packs during development.
How do I add Memory Nexus MCP to my agent?
Install @memory-nexus/mcp from npm, configure stdio MCP in your client, and set the NEXUS_API_KEY environment variable as required by the server manifest.