Secure.
Seamless.
Smart.

Your memory, sealed and yours. It works with every AI tool you already use — and it makes all of them smarter, because they finally remember you.

How it works

Memory, not the model

Most "personal AI" starts from the model and asks you to adapt. MyPenny flips that: it puts your information at the center and lets any AI you trust become more effective. The intelligence is whatever model you bring — MyPenny is the memory that makes it land for one specific person.

Switch models whenever you like. Your memory stays put, portable, and yours.

Three words, one promise

🔒

Safe

Per-user encryption and a public, tamper-evident access log. Your memory is yours to read, export, and delete — always.

⚡

Easy

One connection over MCP works across Claude, ChatGPT, Perplexity, and more. Set up in minutes; remembered everywhere after.

🧠

Powerful

Hybrid search, a learning tag graph, and background consolidation surface the right thing at the right moment.

Safe

Sealed at rest. Watched in the open.

Your memory is encrypted with a key derived for you alone — and every time it is ever unlocked is written to a public, tamper-evident log you can audit. Strong encryption, with real accountability.

  • Per-user encryption. Every note is sealed with AES-256 under a key unique to your account, so a leaked database is unreadable.
  • A public access log. Decryption fails unless it is logged first — you, and any outside auditor, can verify exactly when your memory was touched and why.
  • Yours to edit, export, and delete. Read your memory, correct it, or remove it from your dashboard. No support ticket, no operator in the loop.
  • Honest by design. We don't claim we can't read your memory — we commit that we won't, and the log means you would know if anyone ever did.

Easy

Tell one tool. Every tool remembers.

MyPenny connects over the Model Context Protocol, so the same memory rides with you across Claude, ChatGPT, Perplexity, and anything else that speaks MCP. Set it up once; every tool remembers.

  • One connection, every client. No per-app plugins, and no copy-pasting context between tools.
  • Model-agnostic by design. Bring whatever AI you trust — the memory layer never assumes a provider or locks you to one.
  • Secure to connect, simple to revoke. Authorize each tool with a normal sign-in, and cut any of them off from your dashboard. No API keys to leak.
  • Setup is a conversation. A few natural questions and your assistant already knows who you are and how you like to work — no config files.

Powerful

It finds the right thing — even when you ask imperfectly.

Real memory doesn't make you phrase the query perfectly. MyPenny blends meaning, keywords, and relationships — then keeps learning you in the background.

  • Hybrid recall. Semantic search catches what you meant; keyword search catches the exact term or rare name. A query that misses one still lands on the other.
  • A tag graph that learns your vocabulary. Tags form a graph of parents, children, and synonyms, so a search surfaces related notes you didn't think to ask for.
  • Notes that link like a mind. Memories reference each other, so your assistant can follow a chain of context instead of reading isolated facts.
  • It improves while you sleep. A background pass reviews recent conversations, keeps what matters, and refines your profile — updating notes instead of duplicating them.

How MyPenny works in practice

1

Connect once

Authorize MyPenny in your AI tool with a single sign-in over MCP. No keys to paste, no install to babysit.

2

Just talk

Use your AI normally. Key moments from your conversations become durable memories automatically — no "save this" required.

3

Recall everywhere

Open any other connected tool and it already knows you. Your memory follows you — not your model.

4

It gets sharper

A background pass consolidates what matters and refines your profile, so recall keeps improving the longer it knows you.

Why a dedicated memory layer?

Feature
Built-in chatbot memory
MyPenny
Works across different AI tools
Encrypted with a key that's yours
Export, edit, and delete it yourself
Limited
Recall by meaning, keyword, and relationships
Basic
Improves itself in the background
Open-core engine you can inspect

Honest answers

Is my memory end-to-end encrypted?
No — and we're precise about that. Your content is encrypted at rest with a key unique to you, and every decryption is written to a public, tamper-evident log. We don't claim we can't read your memory; we commit that we won't, and the log means you would know if anyone ever did.
Which AI tools does it work with?
Any tool that speaks the Model Context Protocol. Today that includes Claude on desktop and web, ChatGPT, and Perplexity — with more arriving as MCP adoption grows.
Can I export or delete my data?
You can read, edit, and delete your memory from your dashboard at any time. Full one-click export is on the way; editing and deletion are available now.
Do you use my memory to train models?
No. Your memory exists to serve you across your own tools. The intelligence is whatever LLM you connect — MyPenny is the memory layer, not a model.
Is MyPenny open source?
The core memory engine is built to be open and auditable — designed so the part that holds your data can be inspected rather than taken on faith. A public release is part of the roadmap.

Give your AI a memory that's actually yours.

MyPenny is an early-stage product from HelperLabs, built in the open. Connect once, and every tool you use remembers what matters — only you decide what it keeps.

Get early access