Ori Memory
Give coding agents durable identity, a knowledge graph, and searchable memory across sessions instead of re-explaining context every chat.
Overview
Ori Memory is an MCP server for the Build phase that gives AI agents persistent identity, a knowledge graph, and search over stored memories.
What is this MCP server?
- npm package ori-memory with stdio MCP transport (registry v0.3.0)
- Persistent memory infrastructure: identity, knowledge graph, and search
- Targets AI agents that need continuity beyond a single conversation window
- Open-source repo Ori-Mnemos on GitHub for self-hosted setup
- Local stdio install fits Claude Code and Cursor MCP config patterns
- Published npm package version 0.3.0 in registry metadata
- stdio transport via registryType npm identifier ori-memory
Community signal: 310 GitHub stars.
What problem does it solve?
Every new agent session starts from zero, so solo builders waste time re-uploading context and the agent contradicts earlier decisions.
Who is it for?
Builders running long-horizon Claude Code or Cursor sessions on one product who want graph-backed recall without building memory from scratch.
Skip if: Simple one-shot codegen tasks, teams that already enforce memory solely in Notion or a custom vector DB with no MCP need.
What do I get? / Deliverables
Agents can read and update structured memory so multi-day shipping work stays consistent without manual context dumps.
- Local stdio MCP server wired for agent memory read/write
- Searchable knowledge graph backing multi-session agent work
- Stable agent identity layer across development iterations
Recommended MCP Servers
Journey fit
Persistent agent memory is core infrastructure while you design and implement agent-based products, not a one-off launch tactic. Agent-tooling is the shelf for MCP servers that extend what agents remember and retrieve during development.
How it compares
Agent memory MCP server, not a human-in-the-loop or finance trading integration.
Common Questions / FAQ
Who is Ori Memory for?
Solo builders and agent developers who need cross-session memory, identity, and searchable knowledge for coding agents.
When should I use Ori Memory?
Use it while building agent workflows when context loss between sessions is slowing you down or causing inconsistent plans.
How do I add Ori Memory to my agent?
Install the npm package ori-memory, configure stdio MCP in your client pointing at that package, and enable tools per Ori-Mnemos repository docs.