MemoryAI

One brain. Every AI you use. Forever.

Your AI assistant has no memory. Every session starts from zero — you re-explain your project, your stack, your preferences. Again and again. Tomorrow you switch model, switch IDE, and you do it all over.

MemoryAI fixes that.

One memory. Carried from chat to code to research. Survives every model switch, every IDE, every vendor change. Your knowledge stays yours.

Why install

It remembers. Decisions, preferences, the why behind the choices — all there the moment you need them.
It travels with you. Same memory in your code editor, your chat app, your terminal, your browser. Switch tools, the memory follows.
It stays calm. Long conversations no longer balloon. The window is managed quietly so you can keep talking, not keep trimming.
It's yours. You can take everything with you, anytime. No lock-in to any single AI vendor.

Quick start

Install MemoryAI from the Extensions panel in your IDE.
Click MemoryAI in the status bar.
Paste your key (get one free at https://memoryai.dev).

You're done. The status bar shows when memory is active.

Settings

Open the Connect panel (status bar) or visit IDE Settings → MemoryAI. The defaults are tuned for typical usage; most users only adjust the budget.

Setting	Default	What it controls
Hard cap	150,000	Per-prompt budget
Auto-rotate	on	Manage the window automatically
Recall depth	deep	Quality vs latency for recall
Private mode	off	Local-only for NDA / regulated data
Status bar	savings	Status-bar display format

Privacy

Your key lives in the IDE's encrypted secret storage.
Conversation data leaves the IDE only when MemoryAI needs to persist or recall it.
Turn on Private mode to keep everything local.
Self-host the service — point memoryai.endpoint at your own deployment.

Support

Site: https://memoryai.dev
Issues: https://github.com/memoryai-dev/memoryai-vsx/issues

License

MIT.