Plur
Open standard YAML memory for AI agents — inspect, edit, own.
Finally, agent memory you can actually inspect and edit — no black box, no cloud lock-in. The local YAML approach is refreshingly transparent, though adoption depends on ecosystem growth beyond the listed integrations. If you use multiple AI coding tools, Plur is worth installing now.
- Developers using multiple AI coding agents (Claude Code, Cursor, OpenClaw, Windsurf) wanting unified memory
- Privacy-conscious users who prefer local, inspectable memory over cloud services
- Teams wanting to share agent knowledge via portable Knowledge Packs
- Engineers building custom agent workflows with MCP or Python SDK
- Non-technical users needing a graphical interface or web dashboard
- Users who want fully cloud-managed memory with auto-sync across devices
- Those needing integrations with ChatGPT, Copilot, or other non-MCP agents
We scan live Reddit threads, YouTube comments, X posts, G2 reviews and other communities — and hand you an honest verdict in under a minute.
- Honest verdict, not marketing
- Real pros & cons from real users
- Attributed quotes with receipts
3 free scans · no card needed
In short
Plur — Open standard YAML memory for AI agents — inspect, edit, own. Best for Developers using multiple AI coding agents (Claude Code, Cursor, OpenClaw, Windsurf) wanting unified memory, Privacy-conscious users who prefer local, inspectable memory over cloud services, Teams wanting to share agent knowledge via portable Knowledge Packs. Free to use.
What's new in Plur
Checked 14 days agoAcross the latest 2 updates: 2 feature updates.
Is There an Open Standard for AI Agent Memory Engrams?
No RFC-level standard exists for AI agent memory engrams as of mid-2026; MCP and open engram format are converging.
Should AI Memory Be Stored as Open Engrams or Baked Into Model Weights?
Argues AI agent memory should be stored as open external engrams for inspectability, correctability, deletability, and portability.
Viability Score
How likely is Plur to still be operational in 12 months? Based on 4 signals — momentum (how recently it shipped), wrapper dependency, revenue model, and web presence.
Last calculated: July 2026
How we score →Key Features
- Open, editable YAML memory files (engrams)
- Cross-tool memory persistence (Claude Code, Cursor, OpenClaw, Windsurf, Hermes)
- Local-first, no cloud dependency
- ACT-R decay for automatic memory fading
- Similarity search deduplication
- Structural fix recommendations for recurring corrections
- Knowledge Packs: export/install/share memory bundles
- Multi-store support (personal + team shared engrams)
- MCP server for any MCP-compatible agent
- Python SDK
- OpenClaw plugin (auto-hooks)
- Hermes agent integration
- Lifecycle hooks (init, session start, plan, skill, spawn)
- Privacy scan for secrets and PII in packs
- SHA256 integrity hashes per pack
About Plur
Plur is an open standard and toolkit for persistent, cross-tool AI agent memory. Instead of opaque vector stores or model fine-tuning, Plur stores memory as human-readable YAML files called 'engrams' on your local disk. This lets you read, edit, or delete any piece of knowledge your agent retains, with changes taking effect immediately — no retraining required. Designed for engineering teams using multiple AI coding assistants (Claude Code, Cursor, OpenClaw, Windsurf, Hermes), Plur captures corrections, style preferences, project conventions, and tool routing knowledge, then makes them available across all supported tools without any cloud dependency. Key features include ACT-R decay for automatic fading of stale facts, similarity search deduplication, structural fix recommendations for recurring corrections, and Knowledge Packs for exporting/importing memory bundles with privacy scanning and conflict detection. Plur also provides a Python SDK, MCP server, and OpenClaw plugin, all implementing the same open specification (Apache-2.0, first published March 2026). Performance benchmarks show 97.6% R@5 recall on LongMemEval using a hybrid+reranker approach (fully local), and agents with Plur win 89% of decisive contests and follow house rules 100% of the time (vs 10–38% without). Plur is free and open-source — there is no paid tier or cloud service. The entire system runs locally, ensuring privacy and portability. It is best suited for developers who want transparency and control over their agent's knowledge, especially those working across multiple AI tools. Plur's openness and local-first design differentiate it from proprietary memory services like Mem0, Letta, or Zep, which rely on cloud-hosted databases.
Behind the Verdict
Plur tackles a pain point every AI agent user hits: repeating the same corrections because the agent 'forgets' between sessions. Instead of training a new model or paying for cloud memory, Plur gives you flat YAML files on disk. We think that's the right trade-off for developers who value control over convenience. Pick Plur when you use multiple coding agents (Claude Code, Cursor, OpenClaw, Windsurf) and want a shared memory layer. It's also ideal if you're privacy-conscious — nothing leaves your machine. The benchmarks are solid: 97.6% recall on LongMemEval, 89% win rate in agent contests, 100% house rule compliance. And the cost argument — Haiku+Plur beating Opus alone at 10x lower cost — is compelling. Pass on Plur if you need a managed cloud memory service with sync across devices or a GUI for non-technical teams. Setup requires CLI or MCP configuration; there's no web dashboard. Also, integration coverage is still limited — the listed tools work, but broader ecosystem support (e.g., ChatGPT, Copilot) is absent. Compared to Mem0, Letta, or Zep, Plur's killer differentiator is openness: you own the files, you edit them, you delete them. Those services are cloud-hosted and opaque by default. Plur is also free — but that means no official support, just community and docs. In practice, the ACT-R decay and deduplication keep memory lean without manual cleanup. Knowledge Packs are a neat way to share expertise, though the sharing marketplace is nascent. The real test is whether enough agent tools adopt the Plur spec. For now, it's a promising open alternative in a space dominated by proprietary lock-in.
Researching Plur? Get your full AI stack in 60 seconds.
Free, no signup — tell us your goal and get tools matched to your budget & existing stack.
Use Cases
- Correct an agent's coding style once and have it stick across Claude Code and Cursor
- Store project conventions (tag formats, file placement) so all agents follow them
- Document API quirks and debugging insights as shared team engrams
- Maintain a repository of preferred communication tones across sessions
- Track tool routing preferences among 100+ available tools without re-querying
Limitations
- PLUR is currently a local-first system with limited out-of-the-box cloud sync.
- Advanced users need to manage engram file distribution across machines.
- The ecosystem is early — integrations beyond Claude Code, Cursor, OpenClaw, and Hermes are community-driven.
- No rate limits apply to local use, but the system's effectiveness depends on the quality of user corrections and the diversity of captured memory types.
Integrations
Resources & Guides
Official links
Tools that pair well with Plur
Common stack mates teams adopt alongside Plur, with the specific reason each pairing earns its keep.
Featured Head-to-Head Comparisons
Alternatives to Plur
View allFrequently Asked Questions
Best-of guides
Used Plur? Help shape our editorial sentiment research.