As of 2026-11-21, PLUR 0.23.1 (@plur-ai/cli and @plur-ai/mcp, both requiring Node.js 22 or newer).
Cursor gives you persistent context through Rules: project rules in .cursor/rules, your user rules, team rules, and AGENTS.md files. Those hold instructions you write. PLUR adds the memory that builds up while you work: corrections, preferences and decisions your agent learns, saved as plain YAML in ~/.plur/ on your disk, found again by local search and injected into new chats by Cursor hooks. You can read every memory, correct it, or retire it with one command, and the same store works in Claude Code and Codex.
You need Node.js 22 or newer.
Prefer to let your agent do it? Tell it: "Go to plur.ai and install memory." It follows plur.ai/install.txt: checks for Node.js 22, runs the npm install and plur init --cursor, verifies with plur doctor, and tells you what only you can do (restart).
1. Install the CLI and the MCP server:
npm install -g @plur-ai/cli@latest @plur-ai/mcp@latest
2. Set up Cursor, from your project root:
plur init --cursor
plur doctor
This writes three files into the project:
.cursor/mcp.json: the PLUR MCP server.cursor/hooks.json: hooks that load memory into the chat.cursor/rules/plur-memory.mdc: a rule telling the agent when to save and look up memoriesplur doctor checks the setup and downloads the local search model if it is missing. Without the model, PLUR falls back to keyword search.
3. Restart Cursor.
4. Say yes to memory in this folder. The first time PLUR runs in a folder you have not decided about, it asks once whether to use memory there. Answer in the chat, or run:
plur folders set "$PWD" --on
5. Check it works. Ask the agent to call plur_status. It should report an engram count and the storage path.
Cursor gets a reduced set of PLUR tools: the core memory tools directly, and everything else through plur_admin. Ask the agent to call plur_admin with { "action": "help" } to list them.
PLUR stores engrams: short statements such as a correction, a preference or a convention, each with a scope and a strength that grows with use and fades when unused. It is agent memory, not a code index: Cursor's own codebase indexing still does that job.
A worked example. In a Cursor chat you tell the agent:
"Our API errors always use the
{ error: { code, message } }shape. Don't return bare strings."
The agent saves it with plur_learn. It lands in ~/.plur/engrams.yaml with an id such as ENG-2026-11-17-001. A week later, in a new chat, you ask for a new endpoint and the hooks bring the convention back in.
Read it:
plur list # every engram
plur recall "API error format" # search, the way the agent does
plur dashboard # local, read-only page in your browser
Or open ~/.plur/engrams.yaml in Cursor itself. It is plain YAML.
Correct it. The shape changed:
plur forget ENG-2026-11-17-001 --reason "error shape changed"
plur learn "API errors use RFC 9457 problem+json, not { error: { code, message } }"
To push a memory down without retiring it, rate it: plur feedback ENG-2026-11-17-001 negative.
Delete it. plur forget <id> retires the memory: it is no longer recalled or injected. The entry stays in engrams.yaml, marked status: retired, so you can see what was retired. You can also remove an entry from the file by hand.
To switch memory off for one project: plur folders set ~/client-repo --off.
Cursor's Rules are the right place for instructions that must apply every time, and AGENTS.md is readable by other agents too. Keep using both. PLUR fills a different slot: memory the agent writes as it learns, kept outside any one repository or tool.
Project rules (.cursor/rules) |
User and team rules | AGENTS.md |
PLUR | |
|---|---|---|---|---|
| Who writes it | You | You, or team admins | You | The agent, through PLUR's tools, or you from the CLI |
| Where it lives | The repo, version-controlled | Cursor settings / team dashboard | The repo | ~/.plur/engrams.yaml on your disk |
| When it applies | Always, by file glob, when relevant, or on @-mention | Every chat | Every chat in that directory tree | Memories selected by relevance and injected by hooks |
| Learns from your corrections | No | No | No | Yes, saved as engrams; feedback and decay over time |
| Read and edit | .mdc files |
Cursor settings | Markdown | Plain YAML, plur list, plur dashboard |
| Other tools | Cursor only | Cursor only | Claude Code, Codex and others read it | Claude Code, Codex, OpenCode, OpenClaw, Hermes, Antigravity, any MCP client |
| Other machines | Through git | With your Cursor account or team | Through git | plur sync to a private git remote you choose |
Does Cursor remember previous chats?
Cursor carries context into a new chat through Rules and AGENTS.md, which you write. Cursor's documentation page for its earlier Memories feature now redirects to the Rules page. PLUR adds memory the agent writes itself, which you can read, correct and retire.
Is PLUR set up per project or once?
The Cursor files go in each project, so run plur init --cursor in each project root. The memory store is one: ~/.plur/, shared by every project and tool. Scopes keep project memories apart while your general preferences follow you.
Why does Cursor only show a few PLUR tools?
Cursor gets a reduced tool profile so the tool list stays short. The core tools are direct; ask the agent to call plur_admin with { "action": "help" } for the rest.
Will Claude Code and Codex see what I taught Cursor?
Yes, once PLUR is set up there too (plur init for Claude Code, plur init --codex for Codex). They all read and write the same ~/.plur/ store.
Do my memories leave my machine?
Not by default. The store is local YAML and search runs locally; plur doctor downloads the search model once. Syncing is opt-in: plur sync pushes to a git remote you name, which should be private.
What if the agent remembers something wrong?
Retire it with plur forget <id> and teach the right version with plur learn. For a wider routine, see How to Recover When an AI Agent Remembers the Wrong Thing.
Related: What Memory Layer Works with Cursor or Codex? · How to Test Whether Agent Memory Keeps Projects Separate · Stop Re-Explaining Context: A Restart Checklist for Coding Agents
npm install -g @plur-ai/cli@latest @plur-ai/mcp@latest && plur init --cursor
Open source, Apache-2.0: github.com/plur-ai/plur